Ritu: Hello, listeners. Welcome to Season Seven of the Big Unlock Podcast. My name is Ritu Uberoy, and I’m managing partner at Damo Consulting and your host today. We are welcoming Jean-Claude Saghbini to our podcast. He’s the Chief Technology Officer at Lumeris, where he integrates technological innovations, clinical processes, and operational excellence into leading value-based care solutions. Prior to Lumeris, he has held senior technology leadership roles at Wolters Kluwer Health and Cardinal Health. At Lumeris, JC has been closely involved in developing Tom, the company’s AI-powered primary care as a service platform, and we’re really excited to discuss that today. Welcome to the podcast, JC. Looking forward to an engaging conversation.

Jean-Claude: Absolutely. Thank you for having me. I’m excited.

Ritu: You’ve made the argument that the primary care problem is fundamentally a supply-demand problem, and that adding more digital health applications is not really changing this equation. If AI is going to be different, what exactly needs to change in the operating model? At what point do AI tools stop making physicians ten percent more efficient and actually allow health systems to care meaningfully for more people with the same clinical workforce they have today?

Jean-Claude: Let me unpack that supply-demand problem a bit. Today, about a hundred million people in this country either don’t have access to primary care or have access to extremely inadequate primary care. The reason primary care matters is that it controls about three trillion dollars of spend because it operates so far upstream — and the fact that those people lack access has both healthcare implications for them and significant financial implications at a national level.  The primary reason for the gap is actually simple: there are too few primary care physicians in the country to do the work, and that delta continues to grow. In about eight or nine years, we’ll be approaching a shortage of 90,000 primary care physicians. When you look at it that way, tweaking things and applying AI for small efficiencies is not going to work. The problem isn’t at the scale of five or ten percent — it’s at the scale of a hundred million people. So what we want from AI is technology that can bridge that access gap. With the same infrastructure of physicians and clinicians in primary care that we have today, we need to expand access to the hundred million who aren’t getting it — and to the others who are getting primary care, but only whatever time is available to them, not complete primary care.

Ritu: That’s a great perspective, and we’ve been hearing about healthcare deserts — especially with the rural healthcare transformation project. The problem is even more acute in certain areas. We’ve spoken to an ophthalmologist in Alaska where the wait time can be as much as six months. Tell us more about Tom, which is positioned not as a co-pilot but as an AI-powered member of the care team. Tell us about Tom and how it brings in agentic AI to address this problem.

Jean-Claude: We refer to Tom as two things. Tom is an AI primary care team member, and Tom is also an AI primary care operating system — a platform. I’d love to dig into both, but let’s start with value creation. On the value creation side, Tom is an agentic AI primary care team member that effectively becomes part of the care team of a primary care physician. Just as a physician has staff members to help take care of patients, they add one more: Tom. Tom has access to data about the patient, can compute and reason on that data, identify what to do — from closing care gaps to proactive management. This is what we call the BNA, or Best Next Action engine. It determines the best next action for that patient. But more importantly, Tom can act. Tom can act in a conversational way — we use conversational AI through text messages integrated into the workflow, so Tom can proactively engage with patients and effectively manage them in between visits.  Between visits there is a desert of information. You leave your physician’s office, and the only thing they know about you next is when you show up again. In that time, nobody knows anything about you. Tom not only brings information about you — it actually helps you manage your medical condition between those visits. That is how we’re able, within a fixed human infrastructure, to use agentic capabilities to manage more interactions with patients and a broader set of patients.

Ritu: Tell us a little more about the OS part that you mentioned.

Jean-Claude: To be able to deliver that, you have to build all of the mechanics underneath. The expression of that delivery might be a conversation Tom has with a patient around a cardiometabolic condition they have. But underneath the cover, we had a bit of an unfair advantage: fifteen years of learnings that Lumeris has accumulated on how to manage patients and populations to better outcomes. We’ve partnered with health systems across the country to do that. We took all of that knowledge and embedded it in the operating system.  That gives you components like data ingestion, normalization, and creating a 360-degree view of a patient — whether the data is coming from the EHR, claims, or social determinants of health. We had to build reasoning components that can reason on patients, understand their gaps, and determine what to do about them. We built bidirectional EHR integration — not only to get data from the EHR, but to push data back in. We built patient engagement capabilities: telephony, voice conversational AI, text conversational AI, and a series of agentic services in preventive care, scheduling, chronic condition management, and pre- and post-visit management. We built analytics to surround all of that so you can analyze what’s happening across your population. And last but extremely important — something I’d love to talk more about — we built a safety platform that monitors all of these agents: what they’re doing, tracking clinical metrics, patient sentiment metrics, and objective accomplishment metrics, issuing alerts if things have gone sideways, and ensuring everything stays within the guardrails. The totality of all of that, to us, is Tom.

Ritu: This is really interesting because most of the CIOs we talk to are focused on summarization, documentation, or administrative AI. But you’re saying Tom has agentic capabilities that allow it to make decisions autonomously — some requiring clinician approval, some perhaps not. This raises a lot of questions about trust, governance, and orchestration. How have you approached the trust and governance lens? And when does the human in the loop come into the picture?

Jean-Claude: We use two words internally: guidelines and guardrails, and we live by those. Whenever we build any capability, we define guidelines — what it should do — and then we put the guardrails around what it should not do. We have very strict guardrails: Tom is not a physician, is not a licensed practitioner, and therefore Tom cannot diagnose, treat, or change medications. Those are the guardrails we do not go outside of. Within those guardrails, the guidelines define what Tom can do: engage with patients, check in on them, explain things about conditions, help them manage their condition, check whether they picked up their medication, identify barriers to filling a prescription, connect them back to the care team if gaps are identified, and check on symptoms if a patient expresses them. Those are the guideline capabilities. The guardrails are what we don’t let Tom get out of.

Ritu: Thank you — guidelines and guardrails is a good approach. That leads to the next question: most AI implementations we see have very obvious administrative metrics for ROI. But in this case, where Tom is actually becoming part of the care team, what metrics tell you Tom is improving care and not just creating more interactions?

Jean-Claude: The ultimate metric we look at is access. The definition is straightforward but hard to get to: can the same number of physicians in a health system manage a broader number of patients attributed to primary care? The ROI to the health system is clear — when you have more patients attributed to you, that attribution is monetizable through services the health system provides to that population. The human ROI to patients is that they are now attributed to primary care physicians, whereas otherwise they might wait six or nine months to find one.  On the path to those ROIs, there are first-order KPIs along the way: engagement rate between patients and Tom, care gaps that Tom identified and closed, issues patients raised in conversations that Tom connected to the right venue of care, and symptoms that patients expressed that Tom helped direct to the right place. Those are the first-order KPIs. The second-order KPIs are the ROI outcomes I mentioned.

Ritu: That’s a key point — as long as patients are actually getting access to primary care physicians, you can see the system is working. Let’s change tracks a bit. Tell us about your background and how you got into healthcare. We often get very interesting and personal stories from our guests on that question.

Jean-Claude: I’m an engineer, so I’ve spent my career solving problems. Early on I was in horizontal engineering — technologies across the board. At some point I wanted to get into a vertical I felt passionate about. I was looking at renewable energy as one option, and healthcare as the other. I joke that I got into healthcare because I thought it would be easy. I’ve been in healthcare since 2005 — twenty-one years now.  I got in on the startup side. I’m in Boston, so we started a company in Concord, Massachusetts. In 2005, we had the ambition of creating a SaaS platform to connect medical device manufacturers and health systems and help manage medical devices via RFID from manufacturing all the way through to implantation in patients. That was my entry into healthcare. The technology was very exciting at the time — RFID was the thing back then, the way AI is now. Now it’s ubiquitous. Back then it was early uses of RFID at national scale.  We sold that company — WaveMark — into Cardinal Health. I spent four years at Cardinal Health managing and growing that RFID business from startup scale to enterprise scale. Then I joined Wolters Kluwer Health, where I worked on frontline clinical decision support products like UpToDate, built up the AI Center of Excellence, and did a lot of AI innovation. And this is my third or fourth chapter in healthcare — joining Lumeris about five and a half years ago.

Ritu: Thank you for sharing that story. We have time for one more question. With Tom, you’re bringing together EHR, payer, claims, clinical, and other data to determine the Best Next Action. Do you think the real competitive advantage in healthcare AI will ultimately come from the model — since all we hear about is which model is better or faster — or from the ability to build this longitudinal data and workflow layer?

Jean-Claude: It’s not the models. When any one of the hyperscalers or OpenAI or Anthropic releases a model, eight billion people on the planet have that model the next day. So having that model cannot be a competitive advantage. What I put on my slide — and this comes straight from a conversation with healthcare executives last week — is that the competitive advantage is the data you have. It’s the knowledge you have about care delivery. Many have tried to fix healthcare from a technology lens alone over the past twenty or thirty years, and many have failed doing it purely through tech.  So it’s the data and the understanding of patients. It’s understanding the care delivery models and the domain knowledge — whether that’s primary care, surgery, or cardiovascular. And the third element is building actual applications that leverage these highly democratized LLM models, powering them with your data and your understanding of what you’re trying to do, and then integrating them into workflows — and more importantly, creating new workflows. Integrating into existing workflows is a form of automation, and that’s fine. But now we should be able to create entirely new, reimagined workflows. The totality of that — data, domain knowledge, and applications that create new workflows — is what the competitive advantage actually is.

Ritu: That’s a great answer, and it ties back to what you mentioned earlier — Lumeris had all this proprietary data and deep understanding of the problem that you integrated into Tom. Time has flown by and we’re almost at the end. Would you like to share any closing thoughts or advice for listeners in the healthcare technology field?

Jean-Claude: Three pieces of advice that are top of mind for me. First, don’t fall into the automation trap. There’s an urge to automate things because you can see where AI fits and it’s satisfying to do — but you can get stuck there. The bigger opportunity is reimagining new workflows and new ways of delivering value, not just automating existing ones. I see a lot of pure automation of existing workflows right now. You have to do some of that, but at some point you need to invent the new.  Second, don’t fall into the pilot trap. Pilots were essential in 2023 and 2024 — you cannot learn this stuff from a textbook, you have to do it. But we are in 2026. If you were still just running pilots, that was the top of the funnel. Now we can narrow the funnel, make big bets, and make those enterprise bets.  Third, don’t fall into the perfection trap. LLMs and AI are not perfect, but the problems we’re going after don’t have perfect solutions today either. We can fix enormous problems without perfect AI. The search for perfection will keep technology that patients and clinicians genuinely need locked in a box while we wait for it to be perfect enough to release.

Ritu: That’s great advice. Thank you so much for joining our podcast today. It’s been a wonderful discussion, JC.

Jean-Claude: Absolutely. Thank you so much.

—

Subscribe to our podcast series at www.thebigunlock.com and write us at info@thebigunlock.com    

Disclaimer: This Q&A has been derived from the podcast transcript and has been edited for readability and clarity.

Ritu: Hello, listeners. Welcome to Season Seven of The Big Unlock Podcast. My name is Ritu M. Uberoy, and I’m managing partner at Damo Consulting and your host today. We are very happy to welcome Ann Bilyew to our podcast. Ann is a prominent executive in the healthcare sector, currently serving as Executive Vice President and President of the Healthcare Solutions Group at Internet Brands, where she leads WebMD Ignite, the leading B2B growth partner for healthcare organizations. She has over twenty years of experience investing in and leading healthcare companies, and plays a crucial role in shaping strategies for patient engagement and healthcare solutions. Really happy to have you here with us, Ann. Welcome.

Ann: Thank you, Ritu. It’s a pleasure to be here. Just a couple of sentences about WebMD Ignite so people understand what we do. Most people think of WebMD as the place consumers go with a health question, and that is still very much who we are. We also have Medscape, the leading professional source of information, reference, and news for clinicians. Ignite is our institutional or enterprise business for payers and providers and health systems. What we provide is a comprehensive patient and member engagement platform powered by industry-leading content and data, which enables personalization at scale and omni-channel orchestration. We are really the entire backbone of a patient or member engagement platform for payers and providers.

Ritu: Thank you for telling us about that, Ann. That leads directly to our first question. People are actively searching the internet for health information, and we’ve moved away from “Doctor Google” behavior toward AI summaries — where people don’t even know where the information is coming from; they’re just reading AI-generated answers and acting on them. Patients are now coming to doctors so prepared from ChatGPT that they’re essentially seeking a second opinion on something they’ve already formed a view about. How do we maintain trust in this environment, and how do patients know what to trust?

Ann: A great question, and we don’t really know yet — all of this has happened so quickly. What we do know is that there has been a complete collapse of the traditional funnel for health systems and payers doing traditional digital marketing. By our count, roughly seventy to eighty percent of a health system’s digital budget has been focused on search — specifically Google search. Over the past fifteen to twenty years, they’ve spent billions of dollars trying to be one of the little blue links at the top of a results page. What we’ve seen evolve over the last twelve months is that health is squarely in the crosshairs of AI Overviews — AI-generated answers instead of links. A study by BrightEdge from December of last year found that ninety percent of health-related questions put into Google now return an AI Overview. And when an answer is returned, click-through rates drop dramatically. A Semrush study from around the same period showed that when an AI Overview is present — which, remember, is now ninety percent of health-related searches — click-through rates drop to around one percent. That whole strategy of search engine optimization and search engine marketing has been completely blown up. The information is now readily available whether a consumer goes to Google or uses ChatGPT or Claude or whatever LLM they prefer. What’s increasingly scarce is credibility and trust, and that’s where a health system can maintain relevancy in the marketing funnel. That is where they increasingly need to focus.

Ritu: You’ve made the argument that content is scarce, and AI makes that point so much more consequential. Historically, patient education content could be reviewed, approved, and published. But now an LLM isn’t just giving back that same information — it’s combining it with other sources and generating new answers in real time. How should healthcare organizations treat their content layer in this environment?

Ann: When you look at what used to be called the digital front door — which was always a fancy phrase for a website — traditional SEO was all about keywords, volume of content, and number of pages. Scale. We know that no longer matters as much. Our team studied about fifty markets and well over a hundred and fifty health system websites to see not how they were ranking in Google, but how they were being cited by the LLMs. What we found is that the most important thing you can do is have the right information. When an LLM gives an answer, it is increasingly listing sources — and you want to be cited as one of those sources. To be cited, the content you publish has to be credible: complete, corroborated by third parties where possible, and connected to the overall pathway to care. What we see mattering most is physician profiles that are complete and comprehensive — that really tell a full story about those providers. You need robust information about the service lines that are most important to you, demonstrating your competency and capability, corroborated by third-party sources.  Beyond the citation, what ultimately matters is being listed as the handoff — the place to go for care. It used to be: question in Google, blue link, user lands on your site, and you try to guide them through the patient journey. Health system websites, frankly, are pretty awful. I would encourage any listener who says “my website is great” to actually go on it and see — the friction, the homepage trying to do everything at once between paying a bill, finding a location, and figuring out where to park, none of which has anything to do with getting a person interested in care to the right clinician. Health systems have to own the handoff from question to answer to care. The answers will increasingly be provided by LLMs — full stop. We’re not going to own the information layer. What we can own is the conversion layer, the conversion to care. That is where health systems need to optimize their content, their website, and their outreach.

Ritu: Both of those points are absolutely right. The handoff is so important because information is now coming from so many different sources, and the validity and trustworthiness of citations matters more than ever. Let’s change tracks a bit. We always love an origin story — how did you get into healthcare, how did you get comfortable with tech, and what has your journey looked like?

Ann: I’ve always been in healthcare my whole career — not as a clinician, but trained as a business person. My journey really has three or four chapters. Coming out of business school, I joined a healthcare-only boutique consulting firm called APM, really focused on M&A. This was during the first major transition to integrated delivery networks. Doing M&A, integration, and deep operational work was a great training ground — even post-MBA — to really learn how these organizations operate and how to design solutions that can actually work in the real world.  Then I joined a private equity firm called Advent and became a partner there, investing in healthcare technology and healthcare services, both in the US and internationally — Europe and Israel. I wanted to try running a company, so I took a portfolio company and became CEO of one called MD Everywhere. That’s where I really learned to lead large organizations through periods of rapid technological change. I’ve seen a lot of technological transitions over my career, and a lot of what I call technology theater — most of which ends badly. But this time feels a little different to me in terms of the impact of AI and the velocity of change. It feels different from most other transitions I’ve seen.  We successfully sold and exited that business, and I then joined Internet Brands just after they completed the acquisition of WebMD and took it private. My charge was to build out the institutional or enterprise business — what we now know as Ignite — and scale it. Today we work with 95% of the hundred and fifty largest health systems and 80% of the twenty largest payers, so we’ve gotten a lot done in the last four or five years.

Ritu: Loved hearing those chapters — but still want to go back to the origin. When you came out of business school, why healthcare? There were so many options. Usually people have a personal connection. Is there something that led you there?

Ann: Coming out of undergrad, I was accepted to Harvard Business School — which is unusual right out of college. They said, “You’re young — go work for a couple of years and then come back.” So I had a year to do essentially whatever I wanted. There was a community health center in Kansas City — I went to KU for undergrad — that wanted to build a new facility. They’d outgrown their current one and needed someone to come in on a project basis and help raise the capital through bond initiatives at the city and state level. It was a great project, and that’s really where I got the healthcare bug. I became genuinely intellectually interested in the complexity of the industry and the impact it has on communities and on the country as a whole. We raised the capital, got the building built, and I went back about a year later and saw the groundbreaking — knowing that project had helped many generations of families in Kansas City since then. That’s how I got into healthcare. Coming out of business school, I really only looked at healthcare and went directly to APM — sort of the mothership of healthcare boutique consulting, deploying to all of the leading health systems and payers at the time.

Ritu: Thank you for sharing that. WebMD Ignite describes itself as both a patient engagement organization and a growth partner helping health systems attract patients and strengthen referrals. Most of the time those goals align — but not always. What happens when the commercially optimal action for a health system isn’t necessarily best for the individual patient? And what governance do health systems need to ensure personalization is optimizing for the patient’s best interest rather than just becoming a more sophisticated marketing engine?

Ann: Just a quick clarification first — we work with over 650 health systems in total. The 95% figure refers to the 150 largest, but we work with organizations across the full spectrum: very small, local, regional, and standalone all the way up to the largest health systems in the country. That breadth gives us a tremendous perspective on what’s working, what isn’t, and what strategies are being deployed in different markets.  To answer your question: the way I think about it is that we have to earn the right to communicate with people. The only way to earn that right is to give them something of value — to show them that we know them, understand them, and have taken the time to understand the unique set of circumstances that make them an N of 1. What are their unique clinical characteristics? What are their unique socioeconomic circumstances? What are their preferences — how do they like to receive information, in what format, through what channels? Only when you put all of those things together can you say: here is what matters to this particular individual, here is where they are on their care journey right now. That is how you earn the right to communicate.  This is what retail and other industries have done well and have trained consumers to expect — they want to see that you’ve taken the time to understand their needs, and that you’re giving them information that is relevant, timely, and in their preferred format through their preferred channels. When we give people quality, tailored information, the patient will make the right decision for themselves. That is really our job: give them the right information personalized to their needs, preferences, and circumstances so they can make a good decision — and then facilitate whatever that decision is. Remove the friction from making an appointment, from signing up for a class, offer tools like health risk assessments or calculators. Remove the friction, help people make good decisions. The ultimate decision always belongs to the consumer, to the human, to the patient.

Ritu: What I gather from that answer is that you’ve tried to unify the patient journey. You’ve mentioned removing friction several times, and we know healthcare is deeply siloed. Given that unique perspective of working with 650-plus health systems, what has been the most difficult part of unifying the patient journey, and what goes on behind the scenes?

Ann: I can’t think of another industry that treats its customers — its patients — in such a siloed way. When we think about the patient journey, there’s content as they’re making a decision or selecting a provider or deciding whether to make an appointment — we call that marketing. Then once a patient is on a journey, has selected a provider, made an appointment, scheduled a procedure, the information they start receiving at that point we call education. But it’s all the same thing. It’s all designed to influence someone to make a decision and take an action that is hopefully in their best interest. Whether they’re early in the journey, deciding whether to see a clinician, or they’ve already scheduled a procedure and are receiving information on how to prepare — it’s all information and engagement designed to influence at different stages.  The challenge is that those parts of the organization are completely siloed. Clinical owns patient education; marketing owns patient acquisition and activation; someone else owns retention and adherence. Separate budgets, separate teams, separate technology stacks — yet it’s one person going all the way through that journey. It’s one human being we’re speaking to throughout. Getting our clients to think that way is the work. Some are getting there — you’re starting to see new roles like VP of Patient Experience or Chief Patient Experience Officer, and budgets beginning to come together. But we’re still in the early innings.

Ritu: Time has flown, as usual, and we’re almost at the end of the podcast. Any closing thoughts or crystal ball predictions? We used to ask about the next year or two — now it feels like the next few weeks.

Ann: Anyone making predictions today is taking a big risk. The speed at which we’re evolving as an industry, as a culture, as a country is astronomical and truly unprecedented. I wouldn’t venture any predictions other than this: the information layer will increasingly be owned by the LLMs and the foundational models. Our job is to figure out how to make the connection point between those foundational models and the actual provision of care. That is where we have to concentrate, and that is where we have to figure out how to build.

Ritu: Thank you so much, Ann, for joining us on the Big Unlock Podcast. It’s been our pleasure having you as our guest.

Ann: Thank you. It’s been my pleasure. I appreciate it very much.

—

Subscribe to our podcast series at www.thebigunlock.com and write us at info@thebigunlock.com    

Disclaimer: This Q&A has been derived from the podcast transcript and has been edited for readability and clarity.

Ritu: Hello, listeners. Welcome to Season Seven of the Big Unlock Podcast. My name is Ritu Oberoi, and I’m managing partner at Damo Consulting and your host today along with Rohit. We are thrilled to be welcoming Don Woodlock to our podcast. Don is the President of InterSystems, where he leads the company’s day-to-day operations after spending eight years leading its healthcare business and global healthcare strategy. Prior to InterSystems, Don spent a long time at GE Healthcare. He is also a founding advisory board member of the MIT Generative AI Impact Consortium and hosts Code to Care, a series focused on practical applications of AI and data technologies in healthcare. We’re sure this will be a really interesting conversation — Don brings a unique perspective on how data interoperability, AI, and trusted technology infrastructure will shape the next era of healthcare. Looking forward to it, Don. With that I’ll hand it to Rohit, and then we’ll get started.

Rohit: Welcome, Don, to the podcast. I’m Rohit Mahajan, CEO and managing partner at Damo Consulting and co-host of this podcast. Looking forward to an engaging conversation.

Don: I’m glad to be here. It’s such an exciting era that we’re in — we really need each other in the community to figure out best practices, what’s working and what’s not, so we can direct our energy into what’s best for patients and clinicians everywhere.

Ritu: Let’s get started. You stepped into the president role at InterSystems after spending so much of your career deeply immersed in healthcare technology. As you look across the technology landscape now, what do you think healthcare leaders are still underestimating about the transformation AI is bringing? And what do they really need to do? We’re seeing time horizons shrink from years to months to weeks — it’s almost daily with new models and new headlines. How does anyone keep up, and what should CIOs be thinking about?

Don: In terms of AI, there’s the center everyone talks about — the models, the safety, the applications coming out of the frontier model companies. But there’s a front end and a back end that I think we need to spend more time on. The front end is the use cases, the ROI, the change management. We’re kind of kids in a candy store, and there’s a real danger of eating too much candy — of not being judicious about where we invest our time and everyone else’s. The back end is the data and integration requirements that actually make AI work. It’s not just ChatGPT or Claude applied to healthcare. You need to wire it up, give it data and context, procedures and skills and all kinds of things to make it really perform for your organization. All three legs of that stool need to be attended to for AI to work really well.

Ritu: Great answer. People are really underestimating that. I heard a term recently on LinkedIn — “AI wishing,” as opposed to AI washing — where most people think they can wave a magic wand and AI will just solve all their problems. That leads directly to my next question: healthcare is fundamentally different from other industries where AI agents can simply execute tasks and correct mistakes afterward. In healthcare that simply doesn’t work. Where is the boundary between AI that recommends and AI that acts, when the accountable party always has to be the clinician?

Don: Healthcare is serious business, and AI can’t just be let loose. There are areas — revenue cycle management, administrative scheduling — where you might treat it more like other industries. But for core physician and patient decision-making, AI has to play an empowering role: making sure facts are clear, opening the aperture of what else should be considered. There’s a lot AI can do to support those conversations, but it still has to be human-led. I heard a physician speak recently about a patient who asked what AI thought about their case — but what they really wanted to know was what the physician thought. Patients can figure out AI’s opinion on everything at home. When they come to a healthcare organization, they’re there for human expertise and human touch. Human in the loop is the obvious thing to say, but we need to make it work in a way that truly empowers people. In the end, it’s clinicians and patients who make the choices.

Ritu: I think that’s a great distinction — AI is helping patients prepare for visits and be better informed. But when they get to the doctor, they’re really looking for expertise and reassurance.

Don: From the doctor’s point of view, AI is very good at gathering facts — summarizing a giant chart, surfacing any cardiac history buried in the record, those kinds of things. AI can quickly read a big chart and tell you what’s pertinent. What’s also valuable is the “what else should I be considering?” function — before a clinician zeros in on a diagnosis and treatment plan, AI can open the aperture. What else should I be looking at? The human still makes the choice, but AI is a great input into that choice. It has a lot of positive benefits even from a clinical practice standpoint. But in the end, it has to be used in support of human decision-making.

Rohit: You’re captain of a very large ship that has been sailing for a very long time. From your vantage point, what are some of the things you’re navigating from the AI perspective to ensure your customers and users get the best possible experience?

Don: InterSystems is a software company that supports healthcare as well as other industries, and the software industry itself is going through massive transformation because of AI. So a big part of this is internal — how should AI be used in our own processes to build software, implement it, support it, sell it, market it?  We’ve gone through a few transitions. Step one was basically giving everybody a license to OpenAI or Claude. That gets the wheels turning — people have a personal tool they didn’t have before and can discover how it helps them in whatever role they’re in. Step two, which is where we’re focused now, is embedding generative AI systemically into how we work. It’s no longer a personal choice about whether to use it right now — it’s cemented into the way we operate. For example, when you check in code, it automatically does a code review and posts the results in our system. Nobody is choosing to ask AI to review the code — it’s just built in.  Or take our EHR implementation process outside the US — a year-long program. We’ve built 42 different agents to help employees get their implementation jobs done. One simple example: the first step in an implementation is to read the sales contract and identify what matters to you as an implementer — penalties, specific dates, service level agreements, unusual requirements. We wrote a contract summarization agent. As soon as you post the contract, it automatically summarizes the document — not a generic summary, but one tailored for the implementation person, pulling out implementation-specific items and posting the summary where the rest of the team can benefit from it. That’s embedded right into the work process. The idea is to make the company AI-infused in the way it does its work — that’s the phase we’re in right now.

Rohit: What are you hearing from clients? They’re becoming more well-versed with AI and generative AI, running their own experiments and pilots. You must have a great listening machine over there. What are the common themes bubbling up from the customer side?

Don: There are developer customers and end-user customers. Developer customers are going through a similar transition to ours — generally a little behind us since many are smaller companies or startups. They’re mostly interested in best practices: how should they set up their development environments to best leverage AI on our technology? How do they get our documentation into the context so Claude Code can see it and read it? It’s really about helping them build the best generative AI-assisted development environment with our technology as one of their key platforms.  On the end-user front, it’s use cases, ROI, safety, and governance — all the right issues around how to adopt AI in a way that actually works. And every month the conversation shifts. Right now a big topic is token cost — how do I predict and manage the cost of AI? The implicit question is: will I get an ROI for this use case, or am I spending money without a proportionate benefit? It’s a little like cloud computing a few years ago when cloud costs got out of control. Now people’s token costs are doing the same thing, and they’re trying to figure out how to manage it. The lesson is the same: don’t max tokens — max benefits.

Rohit: The parallel to cloud computing is very pertinent. The same runaway cost dynamic, the same need to get strategic about it.

Don: Both cloud and AI are transformative — but that’s not the same as just letting loose. There’s value in a period of exploration at the beginning, because it’s hard to think of everything when you’re facing a genuinely new transformative technology. But at some point you have to take a strategic view of it, and I think we’re getting into those more strategic views now.  The other big trend, of course, is agentic AI and how it changes the equation for use cases. Healthcare is a little conservative from a technology adoption standpoint, and I think we appreciate that as patients — we want things thought through. In software development, agentic AI is here and now; in healthcare it’s more of a “what could this bring?” question. But I think agentic AI has a bright future in healthcare, particularly for administrative tasks. Imagine telling an agent: “Set up this patient for surgery — get the prior auth, schedule the surgery, check for any transportation needs.” The key is that the agent presents a plan, and the human approves or adjusts before hitting go. That ramps up the importance of human-in-the-loop design and user experience even further. We’re at the beginning of what’s going to be a multi-year era on that front.

Rohit: We’re seeing that directly with one of the startups we work with in AI ophthalmology — the physician is in the loop while the agents work in the background. That’s exactly the model that scales, especially for anything patient-facing.

Don: Healthcare isn’t purely scientific. It involves ethical issues, personal goals and preferences, family considerations — it’s not like software development, which is more straightforward and less wrapped up in medical ethics. AI has a lot to contribute, but it doesn’t carry the risk of wanting to remove humans from the equation. Healthcare is a very human endeavor.  The other dimension specific to healthcare is staffing shortages. Unlike other industries where people are worried about job elimination, healthcare has a genuine shortage of clinicians — physicians, nurses, social workers, physical therapists. People are already waiting too long for appointments. If AI can lower the administrative burden, speed up access to care, and let physicians and nurses go home when their shift ends — those are all very positive things for the industry.

Ritu: Don, would love to hear your thoughts on the AI-native EHR, since InterSystems is also positioning in that direction. With generative AI, ambient interfaces, and agents that can take action — where do you think the EHR is heading in an AI-native world?

Don: Let me take you back to the concept of AI-native for a moment. When generative AI first arrived, some EMR vendors were simply partnering with third parties — bringing in ambient listening technology or chart summarization through partnerships. You partner when something isn’t core to what you’re trying to accomplish. When we looked at it, we said: AI is going to have a huge impact on healthcare, and the last thing we want to do is make it a non-core activity. AI needs to be our competence — built right into the product. Partnering would have cost us six months to market, but what you get when you build it yourself is that it’s yours, it’s core, it’s in the middle of your value proposition. That was the original idea behind AI-native. The rest of the industry has since moved in the same direction, because that’s the right approach.  The goal is always to simplify the workflow for the user. It’s already clunky to use computers when you’re trying to take care of a patient. If you have to jump between multiple software programs, alt-tab, re-log in, and transfer information — that’s clunky. Getting AI right within the workflow, making it a natural part of a unified clinical experience, was important.  I’ll share a story. We implemented a chart summarization assistant — a button at the top of the EMR that opens up the right third of the screen as a chatbot. We visited our first customer who had used it, an ED doctor. I was over his shoulder as he opened a patient, asked four questions of the assistant, then went to see the patient. When he came back, I asked, “Do you ever use the left two-thirds of the screen anymore — the thing we spent fifteen years building?” He said, “No, I never use that.” It had only been live for two weeks. That’s how quickly the world changes. The primary way someone used to learn about a patient was to open the chart, click through the tabs, review the last visit. Now that’s gone.  That’s a read-only use case. But imagine agentic AI delivering more impactful use cases: “Get this patient ready for surgery — do all the steps.” “Work up this patient for this situation.” “Negotiate with the payer’s AI on getting these claims paid.” You state a higher-level intent, and the system figures out what you mean and goes and does it. Software in five years — or less — is going to look completely different. All the user interaction patterns we built over the past twenty years may give way to a new model. That’s what AI-native really means: not bolted on to the last twenty years, but a fundamental rethinking of how end users interact with systems.

Ritu: Great story — I’m sure our listeners loved that. Time has flown by and we’re almost at the end. We usually love to close with an origin story — how did you get into this field, and particularly into healthcare and technology?

Don: When I graduated college I was a math person — I went to MIT, where computer science and electrical engineering were in the same department. I interviewed for both electrical engineering and software jobs. For electrical engineering I pictured myself in a basement with an oscilloscope, wiring things together, and that little circuit would eventually end up in a speakerphone at some hospital — but I would never meet anyone who got value out of what I did. I’d just be part of a long chain. With software, I always imagined I’d be talking directly to end users and building things directly for them. So I took a software job at a healthcare software company — actually an InterSystems customer at the time.  It all turned out exactly as I’d imagined. I remember one of my first projects, in Albany, New York at a healthcare system. I would gather all the users in the morning and ask, “What do you want the software to do?” They’d tell me what enhancements they wanted, I’d code all afternoon, and at 4:00 I’d gather them again to show what I’d built. That direct touch with end users — delivering what they need and seeing them find value in it — has been a core part of my whole career. Fast-forward a few decades, and InterSystems is still a very customer-focused company. I still get to do that all the time, even if I’m not writing production code anymore.

Rohit: Great story. Don, you mentioned the Code to Care series in your introduction — would you like to tell listeners a little more about that?

Don: It’s a video series of five-to-ten-minute videos explaining AI concepts in simple terms for this audience. I enjoy teaching, and I find that AI can be confusing and full of buzzwords. So I started the series to explain things clearly — a recent one covered the shift from prompt engineering to context engineering to harness engineering. Those are terms people hear and don’t quite know what they mean, and I take five minutes to explain them. I have a nice little audience that appreciates learning in bite-sized segments.

Ritu: Thank you so much, Don. Really appreciate you joining us today.

Don: Thank you both. This was a very engaging conversation. I appreciate it.

—

Subscribe to our podcast series at https://thebigunlock.com/ and write us at info@thebigunlock.com    

Disclaimer: This Q&A has been derived from the podcast transcript and has been edited for readability and clarity.

Ritu: Hello, listeners. Welcome to Season Seven of the Big Unlock Podcast. My name is Ritu Oberoi, and I’m managing partner at Damo Consulting and your co-host today along with Rohit. Today we are welcoming Roberta Schwartz to our podcast. Roberta is the Executive Vice President and Chief Innovation Officer of Houston Methodist Hospital in the Texas Medical Center. As EVP, Roberta oversees all operations at this 979-bed hospital, which US News and World Report named the number-one hospital in Texas for fourteen straight years. As CIO, she is responsible for advancing and expanding Houston Methodist’s digital innovation platforms, including telemedicine, AI, and big data. We are really looking forward to this conversation. Welcome to the podcast, Roberta.

Roberta: It is a pleasure to be here. Thank you so much for having me.

Rohit: I’m Rohit Mahajan, managing partner and CEO at Damo Consulting and co-host along with Ritu. Looking forward to chatting with you, Roberta. Over to you.

Ritu: Roberta, you’ve talked about 2026 being the year of maturation and said that healthcare is now well past the AI hype phase. But we’re still seeing CIOs mostly jumping onto a tool-based strategy and acquiring siloed tools — which leads to interoperability problems all over again. With all the success Houston Methodist has had, give us your viewpoint on how AI is actually changing the operating model of a health system, and what you’ve done to make this such a success.

Roberta: If you look at that technology adoption curve that ends in maturation and then starts the next curve, we’re right at that center point. All of us now have fairly mature EMRs that continue to make modifications and improvements — whether you’re on Meditech, Epic, Cerner, or Oracle — layering in more recently available technology: search, symptom checker, chat, texting functionality. That’s the maturation happening in the base tools.  The second thing — and I credit Darren Dworkin as a mentor and friend who articulated this — is that it used to be when we met a company, the job was to keep everyone in their swim lanes so technologies didn’t cross into each other. But that’s no longer possible. What a company could do yesterday, today, and tomorrow are essentially infinite. You used to tell companies, “Stick to your lane,” and now the lane is the entire swimming pool. So picking companies becomes both the challenge and the exciting part — because once you get into the pool with one or two, they can do multiple things with the same technology. The maturation for those of us on the provider side is identifying what companies can actually do, because everyone will sell you that they can do everything. It’s going to be impossible to work with ten companies and keep them each in their lanes. Now you work with one, two, or three that can cover the entire pool for you.

Ritu: That’s a really good answer — especially pertinent today with all the recent announcements from Epic and the other major players. You’re right, nobody’s going to stick to their lanes, and AI is what’s making that possible.

Rohit: Roberta, tell us a little about your background and how you got to where you are today. Listeners usually like to hear that journey — how did you get interested in this space, and what’s been motivating you?

Roberta: I have a 24-year-old daughter who’s currently working for FEMA in Washington, DC and applying for medical school. She’s very focused on disaster medicine — she’s an EMT, got involved with the Texas Department of Emergency Management, and has known exactly what she wanted to do for years. That clarity was partly taken from me, because I also decided what I wanted to do when I was 13 or 14. I started candy striping in hospitals, fell in love with hospitals, and knew I didn’t want to do hands-on medicine. If you’re not going into hands-on medicine, you get involved in the business side. So right out of college I worked at an insurance company for a year, went back for my master’s at Hopkins, then went to work for what is today CMS — Medicare and Medicaid — under the Clinton administration. From there I went into consulting, then in-house at Mount Sinai, and have been at Houston Methodist for 25 years now.  Looking back at that career trajectory, I got into this field to improve healthcare — not from the patient care side, but from the administrative and policy side, recognizing that policy alone could not do everything. My passion for innovation came from watching what innovation was doing in other industries and feeling strongly that it could be brought to healthcare. I’m a late bloomer on the tech side — I’m not even sure I’ve fully bloomed yet. But I understand operations deeply: how you staff, how you change practitioners, what it takes to make things work at the front line. Marrying that with an understanding of what newer technologies are doing and how patients, physicians, and staff are interacting with them — and recognizing that what those technologies did for other industries, they can do for us — that’s the intersection. The issue with many technology companies is they don’t understand the inner workings of how something would have to get approved, what it takes to drive change, or how fragile the healthcare system actually is. You can’t disrupt one piece and assume it won’t impact everything else. But as we chip away carefully, piece by piece, we can make significant improvements in how we deliver care and in the cost of healthcare.

Rohit: That leads naturally to my next question. You mentioned innovation and how you’re embracing it at Houston Methodist. I was fortunate to attend your presentation at the HIMSS AI conference in Boston, and I know Houston Methodist has a defined center of innovation strategy. Would you like to share a little more about that?

Roberta: We began this process years ago. It’s called the Center for Innovation, and we sometimes refer to it as the “digital innovation-obsessed people.” It’s a group of roughly half IT and half operations people who sit together and think through every technology that comes through the door — evaluating the ROI, identifying what it can bring to the front line to improve the clinician’s world and the patient’s world, but also what it returns to the institution. The goal is to give people back time, give people back joy, and give them back the reasons they got into this field. Nobody got into nursing to click buttons or be a slave to an electronic medical record. They got into it to treat patients and feel that sense of accomplishment when they’ve done their job really well. I sometimes get in a little trouble for saying this, but my view is it’s not about ratios and benchmarks. The question is: did you come to work, feel like you did an amazing job, and leave without the anxiety that something was left undone? In these new scenarios, old benchmarks are becoming less relevant — the studies behind today’s benchmarks were done on old technology in different points in time. It’s up to us not only to introduce new technologies, but to push the limits and change the way we deliver care, whether at the bedside, in the home, or in the clinic. And we’re seeing it happen. We’re seeing behaviors change and the way people do things change, and that’s really heartening.

Ritu: At that presentation Rohit shared with me, you described some remarkable numbers around your agentic AI implementation — Houston Methodist achieving a 95% patient satisfaction score for ease of scheduling and generating significant ROI. But as you said, you can’t use the same ratios to measure the clinical side or physician satisfaction. Why have those proved so much harder to quantify, and where do you see the boundary between what AI agents can safely take over and where the human factor still plays an irreplaceable role?

Roberta: Scheduling is a great case study in itself — I could talk about it for a long time, and our leaders Tisha Montgomery and Adam Meyer could talk for hours. But it’s a great example of how technology and change management have to fit together. One without the other doesn’t work. Layering technology onto an old process just costs you more, because you still have the old process and you’ve added the technology on top. Think of the old paradigm where one scheduler serves one or two doctors. If you try to layer technology into that same model, you can’t get any ROI because that person still needs to answer the phone and handle surgical scheduling. You can’t eliminate the role. But when you say, “I’m going to take five schedulers, put them together, and add technology,” suddenly you probably don’t need five schedulers anymore — maybe three, maybe two. That requires change, because people are used to having a dedicated scheduler. But to get from five to two, you first do online scheduling, which requires doctors to put their templates online and keep them accurate. Once that’s in place, there are still phone calls — so who answers them? That’s where our bot, Zoe, comes in. We defined exactly what Zoe handles and when the call drops to a human. Tisha told me a great story where people were calling to say Zoe wasn’t working — it wasn’t that Zoe wasn’t working. Zoe was doing exactly what she was supposed to do, routing clinical questions to the clinic. The issue was the clinic staff weren’t logged in and weren’t picking up. If you don’t have the end of the pipe staffed, it can look like your agentic system is failing when it’s actually working perfectly. Setting up those workflows properly — Zoe gets three tries, then it drops to a human; certain phrases automatically drop to a human — means the human now has more time for the truly complex calls. The combined result of template work, online scheduling, and agentic AI has allowed us to schedule for far more doctors without adding staff. One scheduler can now handle ten doctors instead of four, and a much higher volume is handled online. None of that was possible without rethinking the process first.

Ritu: You’ve made a point we’ve heard from other CIOs as well — the bolt-on approach doesn’t work. You really have to rethink and reevaluate the process before adding AI on top, otherwise you won’t get the transformation you’re expecting. Thank you for explaining that so clearly, Roberta. Rohit, would you like to ask a question?

Rohit: There’s been a huge shift toward virtual medicine at Houston Methodist, Roberta, and it’s especially interesting to us because we’re working with a startup in teleophthalmology as well. Could you tell us how the virtual portfolio has grown over the last five years, and how you’re expanding it with more specialty services?

Roberta: This is without a doubt the direction we’re all heading. We have an enormous number of baby boomers moving through the care system who are older and need more care, while at the same time we have a reduction in the number of practicing physicians in some areas — rural and otherwise — and real access issues in many places. We’re spending a lot of time transferring patients who don’t necessarily need to be transferred, which increases their length of stay, increases resources consumed, and is often unnecessary. Even for Houston Methodist, with eight hospitals ranging from a hundred beds to a thousand beds, each hospital has needs for on-call and consultative services in the moment. Should a patient have to be transferred to our flagship for ophthalmology? The answer is no — you connect with a company that can provide those on-call services. Companies like Genome, which provides genetic counseling on call, with neurological and cardiac genetic counseling specialists — I can’t own ten independent specialists in every subspecialty of genetic counseling. But when we pool our resources, we can each afford a piece of one, and that’s the centralized model. Whether I own the resources or contract for them, the concept is the same: centralize and deliver that care on camera.  We are all in. We have cameras in every room, every operating room, throughout our emergency rooms, and across our clinics. The concept is how do we bring care to you in the moment. I’ll give you a great example even for our current physicians. We got a note from one of our hospitalists who had just left the hospital when he got a new admission. He was picking up his kids, so he said, “I’m just going to camera in.” He did the full assessment, got the patient started, noted he’d camera back in with his findings, and then wrote to us: “It was so great. I didn’t have to drive back half an hour to do that admission. I was able to do it on camera.” That’s our own doctor. Beyond that, we have contracts with multiple providers who help us in areas where we’re short.

Ritu: Switching tracks for the last question — with LLMs and AI agents, performance can shift after deployment, but healthcare has traditionally validated technology only at implementation and assumed it stays static. Do you think this requires a new model of governance? Did Houston Methodist have to rethink that?

Roberta: One of the things I’m most proud of is that when we sat down a few years ago to redo our maturation plan, we said even then: “Be careful. Be wary. Make sure the agents you deploy will not increase costs.” We were already concerned about taking sufficient, baked-in, cost-effective technologies and layering in more expensive ones on top. Agents have the potential both for nirvana and to break the bank. For each one you deploy, the answer is “it depends,” and you have to think it through carefully. Your trials have to be very defined. You have to confirm you’re getting the outcomes and being responsible with the dollars — but that doesn’t mean you should ignore agents or pretend they’re not coming. Go in with an open, skeptical, questioning mind. And always remember: your job is to improve care, improve the workforce, and reduce cost. If you can answer all three of those with a yes for a given agent, move forward and don’t be afraid.

Ritu: We’re almost at the end of the podcast. Would you like to share any closing remarks or lessons learned for our listeners?

Roberta: Change is definitely the hardest part — getting people to understand, to change, to believe. I’m probably one of the greatest evangelists and also one of the greatest skeptics at the same time. I was talking with a group from Rice yesterday and they were excited about something. I told them: “I’m excited about it too. I still don’t think what you’re describing will mature for a number of years, and I don’t think it’s going to deliver the ROI you’re projecting.” Those of us in the field are working with precious few dollars, so anything innovators bring us has to return real value — not “this will bring value if you just get everyone to change and push the whole ball uphill.” That is a very hard ask. My advice: be prepared to really understand the struggles of getting your technology to a point where it returns value, and be honest about where you can and cannot reduce costs for a healthcare system. It’s both an excitement and a warning.

Ritu: Great answer — in a nutshell, ask the right questions and stay questioning. Thank you, Roberta. Thank you so much for joining us today on the podcast.

Roberta: A pleasure. Thank you for having me.

Rohit: Thank you, Roberta.

—

Subscribe to our podcast series at https://thebigunlock.com/ and write us at info@thebigunlock.com    

Disclaimer: This Q&A has been derived from the podcast transcript and has been edited for readability and clarity.

Ritu: Hi, listeners. Welcome to Season Seven of the Big Unlock Podcast. My name is Ritu M Uberoy, and I’m managing partner at Damo Consulting and your host today along with Rohit. Today I’m delighted to welcome Shailu Verma, CEO and co-founder of Mila Health. Shailu has built technology businesses at the intersection of healthcare and AI, serving as Chief Product Officer for UnitedHealth Group’s Strategy and Innovation, leading product at Amazon Web Services, and previously co-founding Landwords, which was acquired by Reuters. At Mila Health, he’s focused on one of healthcare’s biggest operational challenges: how to use AI agents embedded directly into clinical workflows to automate care coordination and improve patient engagement. We are really looking forward to an engaging conversation today. I’ll hand it over to Rohit for a short introduction, and then it’s all yours, Shailu. Thank you for joining us.

Rohit: Thank you, Ritu. I’m Rohit Mahajan, co-host of the Big Unlock Podcast and CEO of Damo Consulting. Very happy to have you on the podcast, Shailu. Over to you for your intro.

Shailu: Thank you, Rohit and Ritu. It’s a pleasure to be here. I’m Shailu Verma, CEO and co-founder of Mila Health. Prior to founding Mila, I worked briefly as a consultant at McKinsey, then started my first company in the computer vision space, which was acquired by Thomson Reuters. I went on to join the AWS AI computer vision team in 2016, building some of the first machine learning and AI APIs that came out of AWS, and then led the machine learning team at UnitedHealth Group. Healthcare has been a deep passion — my wife is a physician, I have two daughters, one in medical school and one working toward getting in. Getting to apply technology in healthcare has been a dream, and that’s how Mila Health was founded.

Ritu: Most of the healthcare industry still thinks of patient engagement as reminders and messaging — more apps that patients have to use. But you’ve argued that the real opportunity is AI-driven care coordination that actually completes clinical workflows and helps the patient. What does that shift look like in practice, and why is it so much more impactful than traditional outreach? And this wouldn’t have been possible two or three years ago — but now, with AI, we finally have the tools. What does Mila Health see as the real opportunity here?

Shailu: Let me go a little deeper into the problem. Healthcare organizations have patient engagement tools, EMRs, and many different software systems that they’ve invested in over time, all connected in a complex web. But each platform follows its swim lane very narrowly — patient engagement does reminders, EMRs handle records, payment systems handle billing. The aspect of actually taking care of a patient through their care journey is still done mostly manually. Reaching out to bring someone in for an annual wellness visit, preparing them for a procedure like a colonoscopy, guiding them on medications and diet, following up afterward to determine whether they need another procedure in five or ten years — all of that is still largely manual work. It’s a $334 billion problem. Connecting these complex systems and closing the gaps in patient care was a hard problem a few years ago. Now, with large language models, we have the opportunity to reduce that enormous manual effort in giving patients access to care and coordinating care after appointments. That is Mila’s mission — and to answer your question directly, while patient engagement tools address reminders, there are still many gaps in full care coordination, and that’s what Mila is solving for.

Ritu: You’ve said that organizations moving fastest with AI are not just buying another point solution — they’re embedding intelligence into the systems clinicians already use. Why is that distinction so important, and what have you learned from driving adoption inside health systems? How do you get them to adopt yet another solution when they already have a whole range of apps and software?

Shailu: To automate the vast majority of patient care journeys that are done manually, and to make it easy for our customers, we knew we had to build systems that get embedded into existing software. We took inspiration from companies like AWS, Stripe, and Twilio — they have APIs that get embedded into an existing stack and are easily configurable. A health system shouldn’t have to reconnect Mila to their EMR; it should come embedded and connected through APIs. We built Mila to be completely embedded — you can turn it on in days or weeks rather than months of complex integration. That makes it part of an existing solution rather than another point solution. Point solutions have two problems: implementation time, and they become yet another swim lane. We want to be the interconnectivity between swim lanes, not create a new one.  We built Mila as a set of APIs that any existing software system can adopt to automate across other systems. To give you a concrete example: we are live in over three hundred dental organizations in the US. Each has an EMR, a practice management system, a payment system, and an insurance verification system. Our APIs are embedded in the practice management system those organizations already use. To adopt AI for care coordination, all they have to do is flip a switch in their existing system and say, “I want AI to reach out to patients who haven’t had their cleaning in the last year.” No separate EMR integration, no practice management integration, no telephony integration required — it’s already built in because the APIs are embedded in the stack. That’s how an API-enabled, easily configurable solution scales effectively and makes it easy for customers to adopt.

Ritu: We’ve heard repeatedly from C-suite leaders that healthcare has no shortage of impressive AI pilots, but very few reach enterprise scale. You’ve described Mila now supporting hundreds of organizations because of ease of integration. What were some of the other key architectural, clinical, or organizational decisions that enabled you to move beyond pilots to real operational deployment?

Shailu: Our enterprise scale has happened primarily by making it simple for organizations to adopt — because it’s embedded in their software stack and almost invisible, the way Twilio is invisible in software today. And we address problems that are critical to their immediate needs. Take the dental organizations I mentioned — they’re short-staffed, unable to consistently reach their patient panel or follow up effectively after care. Many calls go unanswered, hit answering machines, or come in over weekends when patients need help immediately. We embedded Mila into the practice management stack they already use and solved the real problem: answering patient questions 24/7. Patients can call any time, change or create appointments, ask questions. We also built in a human-in-the-loop mechanism — when Mila can’t answer, a human steps in. That makes it a safe, easily implemented solution, and that’s what enables enterprise scale versus ongoing pilots where you have to prove things over and over. The formula is simple: make it simple, solve real problems, and bring a human in the loop when needed.

Rohit: As you scaled the business, what were some of the core challenges you faced — both business and technical — especially given the wide variety of practices you work with, from small to large? How do you handle the integrations?

Shailu: Building a startup is hard, and building one in healthcare is even more complex. The challenges are mostly universal startup challenges — get more customers, get more funding, build the product, keep the team motivated. Getting that virtuous cycle moving is the core of the startup journey. On the technical side, three tenets that we established early have been foundational. The first was to build as an API-first product — white-labeled, easily adopted by any software system, easily configured. Not a point solution. The second was configurability. Every practice, small or large, has its own rules — their own three-inch binder of processes. Mila needs to adopt that binder for each office and adapt accordingly. The third was safety and governability. Mila must be consistent — if a patient asks the same question five different ways, Mila should give the same answer each time. It should be safe. And a health system of any size should be able to govern it and understand why Mila said what it said. We established these three tenets before we wrote a single line of code, and they have held true to where we are today. That is one of the key reasons we’ve achieved the penetration and enterprise scale we have — it makes adoption easier, governance easier, and change management easier for our customers.

Rohit: What are some of the new things you’re working on? Products in SaaS are always evolving based on customer feedback. What’s bubbling up to the top? And I’m curious about the fast-changing LLM landscape — with Chinese models emerging, cost structures shifting. How do you navigate that?

Shailu: There’s a common thread to both questions. On the technical side, we’ve built a neurosymbolic graph that sits on top of large language models. We are LLM-agnostic — we swap models regularly depending on the task, optimizing for cost and efficacy. Our thesis has always been that in healthcare, you cannot let an LLM use a stochastic bag of words to answer a patient’s question. The health system needs to be able to configure and control what Mila says, and there needs to be explainability for why Mila said what it said. With that architecture, we welcome the changing landscape — foundational models getting better and cheaper. Token costs have already come down a hundredfold, and we think they’ll come down another hundredfold. But the last-mile problem of governance, consistency, and configuration remains, and that’s where Mila has innovated. That’s also where our ongoing development is focused: moving from conversation to autonomous execution of tasks. It’s one thing to have a ChatGPT-style back-and-forth. What matters for our customers is execution. We’re working with a large payer right now where we’re calling Medicare Advantage members to schedule their first annual wellness visit appointment. Previously, the payer would spend three months preparing with a call center and achieve an outreach rate of fourteen to eighteen percent at a cost of twelve to fifteen dollars per patient. With Mila, we’ve reduced startup time from three months to six weeks, increased the response rate from fifteen to eighteen percent to over fifty percent, and reduced cost per outreach to one-fifth of what it was — all while having a conversation with the patient and actually executing the appointment scheduling. That last-mile focus on efficiency and outcomes is where we’re doubling down.

Ritu: Many CEOs and C-suite leaders are feeling pressure to do something with AI but are unsure where to begin. Based on what you’ve seen across your customers, what’s the biggest misconception about deploying AI in healthcare production today? And where would you recommend they start to generate real, measurable ROI?

Shailu: The place where I’ve seen success happen most consistently is where AI solves a revenue problem or a patient intake problem. Every provider in a health system has thousands of patients in their panel. A GI doctor may have twenty thousand; a primary care physician three to five thousand. If you ask almost any of them, they’ll say forty to sixty percent of their patients haven’t been seen often enough or haven’t completed things like their annual wellness visit. The problem is twofold: patients aren’t being seen, and health systems have quality and prevention metrics — colonoscopy screening rates, cancer screening rates, annual wellness visit rates — that they need to hit. And of course, each visit generates revenue. We’ve seen real success when we address the simple problem of reaching out to this disconnected patient population, activating them, and scheduling their appointment on a single outreach call. That solves the metrics problem, addresses the patient care gap, and generates meaningful revenue. A concern I hear from leaders is: what will patients think of an AI agent calling them? We’ve now done this millions of times. Roughly three to four percent of patients say they don’t want to speak with an AI — in which case, a human steps in. Some hang up. But the vast majority, ninety to ninety-five percent, are genuinely happy to hear from their care team and learn they can schedule a visit right then on that call. That is real change. And once they’re back in the system, everything else follows naturally — follow-up care, ongoing coordination. That frictionless starting point is the most powerful place to begin.

Rohit: Any thoughts about the future or closing remarks for the audience?

Shailu: Watch for the conversational part of AI to move from conversation to execution — to actually getting work done. That’s super important for Mila, but it should be equally important for healthcare leaders trying to implement AI effectively. And people should not be integration layers. The vast majority of the manual work happening today is integration work — verifying insurance, closing scheduling gaps, following up on outstanding items. These are simple gaps that AI can close. We don’t need humans calling in to update an insurance record. Let’s start simple, focus on execution, and use the tools we have. It isn’t rocket science. We’ll solve bigger problems over time, but these easy wins are the right place to begin.

Rohit: Great thoughts, Shailu. Thank you for joining the podcast. Over to you, Ritu, for closing.

Ritu: Thank you, Shailu. It’s been an absolute pleasure. Thank you for sharing your insights, and I hope our listeners take away a great deal from today’s conversation.

Shailu: My pleasure. Thank you for having me.

—

Subscribe to our podcast series at https://thebigunlock.com/ and write us at info@thebigunlock.com    

Disclaimer: This Q&A has been derived from the podcast transcript and has been edited for readability and clarity.

Ritu: Hello, listeners. Welcome to Season Seven of the Big Unlock Podcast. My name is Ritu Uberoy, and I’m managing partner at Damo Consulting and your host today. We are absolutely thrilled to welcome Dr. Ami Bhatt to our podcast. Dr. Bhatt is a cardiologist and the Chief Innovation Officer at the American College of Cardiology, and Chair of the FDA Digital Health Advisory Committee — two roles that place her at the center of how healthcare institutions adopt, govern, and scale technology. A Harvard and Yale-trained cardiologist, formerly at Mass General, Dr. Bhatt has spent her career inside the systems that deliver care, which is precisely why she understands where and why they break. Her work spans AI governance, digital health strategy, and health equity, and we could not be more thrilled to have her here. Welcome, Dr. Bhatt.

Dr. Bhatt: Thank you so much for having me. I really appreciate it.

Ritu: We like to ask our guests about their origin story — how they got into medicine, healthcare, and technology. Would you like to share something?

Dr. Bhatt: Like every first-generation kid in the US, I liked science, I liked people, I went into medicine — that’s what I always thought my story was. It wasn’t until I took the time to think a little deeper that I realized what I love about medicine. In no other profession do you get entered that deeply into somebody’s life within fifteen minutes of meeting them — not just their health, but their family, their hopes, what they want. That connection has been the driving force behind why I became a clinician and why I stayed one for so many years.  I practice adult congenital heart disease — a young population. Kids who grew up and came to me at 16 or 18, and those teenagers became my patients into their 20s, 30s, 40s, and 50s. They were living with lifelong disease. Coming to the hospital wasn’t a one-time thing — it was a reminder of a medicalized childhood. I had patients who said the smell of the hospital made them anxious. So about seven years before COVID, I started doing telemedicine for my patients. At the time, not many people were doing it, and colleagues would ask, “Dr. Bhatt, how are you going to take care of these complicated patients without touching them?” I started exploring what digital technologies I needed — digital stethoscopes, handheld ultrasound, image sharing — and how to teach patients to feel empowered and understand what they needed to do.  I had a patient once up in Maine whose local doctor said something was wrong with his heart because he had passed out at softball. He said, “No, I tripped — and my heart always sounds like that.” He played a recording of his heart murmur that I had recorded in 2014 with one of the first digital stethoscopes. I had told him: “If anyone ever tells you this is a problem, you play that recording.” That patient felt genuinely empowered. That model — staying at home, taking visits remotely, teaching local providers how to care for them, while knowing I was there when they needed me — carried with me through outpatient cardiology at Mass General, through COVID, and into how I think about innovation today. How do we empower patients in the community with wearables and other tools to understand their own baseline, and when a trigger or a change appears, know it’s time to escalate to clinical care? That’s how I ended up where I am today.

Ritu: Thank you, Dr. Bhatt. Starting telecardiology seven years before COVID — you were truly a pioneer. I was reading your work, and you’ve written that clinical acumen supersedes AI every time. That’s a powerful statement in this era when AI models are increasingly capable. You’ve also said that clinicians are not short on intelligence or judgment. So how do we build healthcare systems that benefit from AI while ensuring clinicians don’t lose the observational skills, intuition, and pattern recognition they’ve spent years developing?

Dr. Bhatt: Humans are short on compute power — that’s what we are. The amount of information available to us now far exceeds what the human brain can utilize. When you’re sitting in front of me, I have access to not just your health records but your personal and family history, your preferences, your social determinants, your wearable data, your labs, your imaging, the research on the multiple diseases you may have by age 60, and novel findings that haven’t yet made it into the guidelines. How can I process all of that in twenty minutes? I cannot. That is the role of AI: to surface knowledge, to surface information, and to use compute power to see relationships and trends that would take us a very long time to identify. Back in the day, there was less information and more time. One chapter about your disease. Just the patient in front of you. No EHR. An hour to ask the same question again and get a full story. Nowadays, who has an hour with a patient? We understand why colleagues go into concierge medicine — because that intimate connection is the magic of caring for patients. In its absence, we need compute power to get there faster, and then use that information and those insights to ask: does this fit this patient? Let me use my clinical acumen to decide what happens next. That’s the goal.  I always tell my fellows: you have to use AI until your brain hurts. A recent paper I read yesterday showed that knowledge workers actually spend more time, not less, because of AI — asking questions, interacting, refining. Most of us have accepted that AI is here. We see the numbers. But now the challenge is understanding when to use it. When should you deep dive with it? When does using it take time away from a patient conversation? We haven’t fully studied that yet. At the American College of Cardiology, one of our goals is to have our own members using AI for question-and-answer tasks and understand: how do they know when they’re getting better results? How can they recognize when they’re being led astray? We are at the very beginning of understanding how humans and AI will interact in clinical medicine.

Ritu: Context comes up repeatedly in your recent work. You’ve argued that medicine is not just about recognizing patterns but about understanding the patient’s lived experience and the clinician’s accumulated experience. As healthcare moves toward increasingly autonomous AI systems, how do we ensure that context becomes a design requirement rather than an afterthought?

Dr. Bhatt: The first thing I think about is starting with the patient. The patient is not given the level of respect they deserve in this equation right now — though we’re moving toward it. Rather than me trying to figure out your social determinants, why not have you bring that information directly to me? How do we capture a patient’s feelings, tendencies, preferences, and ways of engaging with healthcare? What is best for this individual, and how should they be engaged? We ask how a doctor likes their notes written, but we don’t ask how a patient likes to be engaged. We do a little of it — do you prefer telehealth or not? — but there’s so much more depth to that question. We have to start by engaging patients and helping them understand how they want to interact with the system, express it, quantify it, and then communicate it so the system can respond. That’s number one.  Number two is around agentic AI, and it’s actually one reason agentic AI can sometimes be better — because I can always tell my AI: say this at an eighth-grade level, I don’t like that tone, say it differently. I can’t say that to a colleague sitting across from me. My patients actually did this all the time. Young patients in their 20s and 30s would say, “I didn’t understand that, Dr. Bhatt. Can you draw it again? Your drawing isn’t very good — can we try something else?” Honest, direct feedback. I want that for every patient. Agentic AI can provide that, sometimes better than a rushed clinician. Any clinician given enough time will do all of that — we’re just not given the time.  The second thing in agentic systems is where the clinician places the guardrails. I wrote about this in my most recent Substack: in cardiology, there are specific places where AI should not make a decision without a human. We have to define those hard stops for the companies or systems implementing this. But beyond that, as agentic AI is being built, we should be able to give our input case by case. I was just talking this morning about a company called UpDoc that is trying to take clinician input as agentic AI is built and let individual doctors set their own guardrails — telling the system what they’re comfortable with and not comfortable with. That’s a lot to manage, but I think those are the right things to study: how to help agentic AI make the right decision at the right time, with the right guardrails, so physicians feel in charge — just as we want patients to feel in charge of their own process.

Ritu: You’re also challenging the industry to think beyond whether AI can generate the answer, and asking what happens after that answer reaches the patient. As health systems deploy more AI-generated insights, do you think the next frontier is actually communication — helping patients understand, trust, and act on these recommendations?

Dr. Bhatt: It’s already communication. Look at the wearable industry. The reason people across all socioeconomic levels, educational backgrounds, and geographies are increasingly wearing some form of wearable is that they want to engage, learn, and find data about themselves. And the reason it works isn’t only that the person wants it — it’s that the systems these apps create are inherently engaging. They know how to make a patient feel empowered. That’s already happening. What I’m focused on is how we take that industry and bring it into the flow of clinical care — and back out on the other end. We lose people everywhere in clinical care. Technology companies have a responsibility to come with us on the whole journey: “I see a signal. We’re going to provide some care. Now they need hospital care. They’re back out — and you’re back in the system that caught you in the first place.” That’s what we’re trying to build.

Ritu: Speaking of agentic systems, tell us more about the ARPA-H Advocate initiative, because they’re using agents and agentic AI to do exactly what you’re describing — continuously monitoring and intervening. What are your thoughts on that program?

Dr. Bhatt: Haider Warraich is a friend — he’s a cardiologist here in Boston, a heart failure physician. With ARPA-H, you can apply with your own idea and say, “Can I have funding to make this happen?” And that’s what he did. What I like about this approach is that it’s a pure research approach. The goal is to prove that you can have an agentic system that works, that can risk-stratify, that knows where a human in the loop is necessary, and that can monitor what happens as it goes along. The key problem we have right now is: we implement something — digital health, AI — and then it’s out there, and we don’t know what will happen afterward, what will change. We don’t have an infrastructure for post-deployment monitoring. What I’m most excited about with Advocate is that we’re going to have concrete examples of what post-deployment monitoring looks like — how we sense trends and shifts, how we decide to iterate and adapt. That’s the piece we’ve been missing. How do we keep track of what’s happening after we deploy?

Ritu: Dr. Warraich was actually on our podcast a couple of weeks ago talking about ARPA-H and the Advocate program — a very interesting conversation. Tell us more about your work on the FDA Digital Health Advisory Committee. This comes up constantly in our conversations with CMIOs and CIOs — things are moving too fast for regulation to keep up. By the time you get FDA approval, the underlying model has already changed. How do you see this gap closing?

Dr. Bhatt: I have the utmost respect for my colleagues at the FDA who are there full time. As inaugural chair of the Digital Health Advisory Committee, one thing became very clear over those two years: there is a real need to convene patient groups, industry technologists, and clinicians together to understand how the FDA can create an infrastructure for care — not simply regulate or deregulate, which is where we tend to get stuck. We need to understand from all of those perspectives what guardrails are needed, what the technology companies need, and what fits the clinician’s workflow. That unsiloing is what’s happening at the FDA right now. You’ll find them at every table, in every room, maintaining their commitment to scientific rigor. FDA approval is the table stakes — it means your technology works. But we’re now thinking about what comes next: not just post-deployment monitoring, but a pre-control change plan. With a medical device like a valve, you can tell the FDA what’s in the pipeline. With AI, I don’t necessarily know what model changes are coming. So how do we use real-world evidence to understand where things are going and how they’re changing?  The FDA did allow some trend measurements and monitoring capabilities in the cardiometabolic space to proceed without formal approval, which was announced earlier this year. That creates real-world evidence about how quickly you can adapt from community signals — sick or not sick, at baseline or changing — and route patients into the clinical arena accordingly. Agentic systems can handle the lowest risk; high risk clearly needs a human; we’ll be in the middle ground for a while. I’ve learned that real-world evidence is essential. Everyone is now at the table, including the FDA, which I love. But we’ve never had a time when change itself is the constant — not just the technology, but the uptake, the expectations of the clinical network, and the demands of patients. All of that is changing simultaneously. All we can do is keep working together, look for real-world evidence, and wherever we can find data, value it and study it.

Ritu: Tell us a little more about the Elevate Leadership Program at Mass General — with everything moving so fast, clinicians need to be up to speed, but you’re all already so busy. How do you find the time?

Dr. Bhatt: About four years ago, a grateful patient family foundation was willing to give $3 million to Mass General Hospital to fund a new style of leadership course. The course I co-wrote with my colleague Eytan Shapira from MIT was designed for leadership-level clinicians — cohorts of fifteen — who were about to lead large groups of people. The longstanding problem in healthcare is that the better you do as a clinician and researcher, the more likely you are to receive an administrative or managerial role. Those skillsets don’t automatically overlap. They can, but they need to be taught. There have been many leadership programs for senior people, but I wrote this one because I saw the technology change coming — I had already been in it for ten years. COVID was the culmination of many of us in digital health saying for a decade: high-quality care in the community where people live, tech is coming and it’s going to come faster. We weren’t ready.  The course works on how you first evaluate yourself as a leader and understand the discomfort that comes with change — how to express and share that discomfort while still being seen as a leader, and how to look at what’s coming and identify the true unmet needs versus the shiny things that aren’t necessary right now. We’re also reversing a long-standing dynamic in academic medicine where clinicians kept industry at arm’s length. Now we’re saying: let’s talk about what technology you have, because if you don’t hear from me what I need and how it would fit, you’ll build something that doesn’t get used. Eighty-six percent of cardiovascular digital health and AI technologies that pass the FDA sit on the shelf — they never get deployed. Passing the FDA does not mean success.  The goal of Elevate is to create a community of leaders who lead not only with empathy but with an understanding that discomfort is constant — it’s not going away, and it’s because of constant change. I am here with you, respecting that, understanding the challenges, and helping find the right next step to move forward. That was my last project before I left MGH, and I’m really proud of it.

Ritu: What you said is so valuable — change is inevitable, and you really need to face it. Preparing people to sit with discomfort, to accept that they won’t know everything, but that as long as they keep learning they can navigate it.

Dr. Bhatt: That’s exactly right. And change is no longer in the future — it’s right now. If you run a practice today, technology is coming at you out of a fire hose, constantly. If you are a healthcare leader in the next decade, there will never be a moment where things feel stable. They will always be bubbling. You just have to decide: when am I doing things, and how will I act despite constant change?

Ritu: Thank you, Dr. Bhatt. Would you like to share any closing thoughts with our listeners?

Dr. Bhatt: The most important conversations happening right now are those where technology, clinicians, and patients are put in the same room to think about what the systems of care should actually look like. I would love for every listener — whether you’re a venture capitalist, a clinician, or a patient — to ask: what is the infrastructure for care in my world? Do we have one? Do we have guardrails? If not, who do I need to talk to in order to put something in place? We’ll keep working on this from the national level at the ACC and through the technology companies I work with, but I think each individual taking a moment to ask “is there an infrastructure, or is healthcare just hard because we don’t have the right one?” is powerful. The infrastructure we have was built in the 1970s, and that is not today’s healthcare. Look for infrastructure — it sounds boring, but it’s everything. And if people want to reach out, I’m Dr. Ami Bhatt on LinkedIn and Substack — I’m happy to talk and hear more ideas.

Ritu: Thank you so much, Dr. Bhatt. Lovely talking to you, and I’m sure our listeners will take a great deal away from this conversation.

Dr. Bhatt: Thank you for having me.

—

Subscribe to our podcast series at www.thebigunlock.com and write us at info@thebigunlock.com    

Disclaimer: This Q&A has been derived from the podcast transcript and has been edited for readability and clarity.

Ritu: Hello, listeners. Welcome to Season Seven of the Big Unlock Podcast. My name is Ritu Oberoi, and I’m managing partner at Damo Consulting and your host today. We are welcoming Michael Archuleta to our podcast. Michael is the Chief Information Officer at Mount San Rafael Hospital, where he has helped transform a rural critical access hospital into a nationally recognized leader in digital innovation, cybersecurity, and healthcare IT. Michael is a perennial Becker’s Hospital CIO to Know, nationally recognized speaker, advisor to health technology companies, and someone whose work has spanned cloud transformation, cybersecurity, and AI adoption — and more recently, one of the most significant technology initiatives in the hospital’s history: the transition to Oracle Health. We are really excited to have Michael here with us today. Thank you for joining us on the podcast.

Michael: Thank you very much. It’s definitely a pleasure to be here.

Ritu: Michael, would you like to tell us a little more about Mount San Rafael Hospital so our listeners have some context?

Michael: Mount San Rafael Hospital is located in Trinidad, Colorado. Las Animas County is actually the largest county in the state of Colorado, with a population of about nine thousand in the city and around fourteen thousand across the county as a whole. I’ve always had a passion for rural healthcare, and I’ve always stated that your zip code should never determine your healthcare outcomes — that is such a critical element. Here at Mount San Rafael, we believe that whether a patient comes to us, to Denver Health, or to UC Health, they should receive the same level of care. We focus heavily on innovation and technology enhancements because in rural America, unfortunately, zip code does still determine outcomes. But we’ve been fortunate to bring a lot of innovation to what we do, and we’re proud to be a nationally recognized rural healthcare hospital within the state of Colorado.

Ritu: You’ve described your Oracle Health implementation as much more than an EHR replacement, and you recently shared the principle of “follow the patient journey, not the org chart” — a really powerful leadership philosophy. Tell us more about this project, and what advice would you give other CIOs embarking on similar modernization efforts?

Michael: I’ve always said that if you work in healthcare, your new CEO is the patient — that’s the bottom line. As a technology and business leader, the goal is to build tools that benefit our patients both inside and outside the organization. Patient care is number one. We have to look at initiatives that enhance patient care outcomes, improve the way we share information, and give clinicians a complete view of the patient’s health — along with the right toolsets to drive better outcomes. AI, used as an enhancement tool, can create better outcomes, better automation, better operations, and better clinical documentation. We put specific AI algorithms into our radiology practice, and I can literally say those have been life-saving. When an organization is considering replacing a system or investing in something new, you have to ask: what’s the why? Is it better operations? Better patient care? Every digital investment we’ve made has had a patient-centric strategy attached to it, because that’s how critical it is. Hospitals and clinics are digital companies that deliver healthcare services — we have to move away from the brick-and-mortar theology healthcare has been built on. Our strategic partnership with Oracle Health allows us to move into the next era, utilizing AI-native technology built directly into the process itself, from the EHR to the patient portal. That is our mission moving forward.

Ritu: We had Dr. Bharat Sutariya on the podcast from Oracle Health, and he also talked about being AI-native and how that’s going to be a real game changer — having everything built in rather than increasing clicks. I think you’ve leapfrogged a lot of health systems by moving ahead with this implementation.

Michael: A hundred percent. This move to Oracle Health was a strategic investment in the future of this organization. By integrating true clinical AI with a modern, unified EHR system, we’re giving our caregivers technology that removes barriers. How do we remove those consistent clicks? How do we let providers be providers again, instead of being attached to a workstation? Incorporating ambient AI removes those barriers, enhances clinical documentation, and reduces burnout. Providers are constantly seeing patients and sometimes fall behind on documentation — doing it after hours or at home. We have to show we’re investing in technology to create better automation, better efficiencies, and reduced administrative strain, while strengthening the connection between physicians and patients. This positions our facility to deliver a high-performing, more connected, and more sustainable model of care for the region.

Ritu: With AI becoming embedded across clinical workflows, you talk a lot about cybersecurity as fundamentally a patient safety issue rather than just an IT responsibility. How do CIOs need to approach cyber resilience as an essential component of clinical quality rather than just operational risk?

Michael: Cybersecurity has to be — and I consistently repeat this — a core component of the overall organizational strategy, period. I was looking at some data points from a national survey sent to many health systems, and only about 43% of healthcare organizations provide basic cybersecurity awareness training to their end users. We have a major gap where organizations aren’t focusing enough on security, and healthcare has gone from being in the top ten to the most attacked industry. When I go to my board of directors, I focus on three elements — what I call FOR: Financial, Operational, and Reputational. If a cybersecurity breach happens, what are the financial consequences, the operational consequences, and the reputational consequences? Framing it that way tells the story of why we continue to invest in security. Deep fake technology, especially with AI, is becoming more advanced than ever — voice clips, video clips designed to look legitimate. You’re on a Zoom call and you genuinely don’t know if you’re speaking with a real person or a deep fake created by a threat actor trying to compromise your organization. Social engineering is equally critical. We did something a bit out of the norm: beyond penetration testing, policies, and procedures, we ran a social engineering exercise where an individual with a badge similar to ours tried to access the facility to see how much access they could gain across different departments. It opened everyone’s eyes — attentiveness and vigilance are extremely important, and cybersecurity can be physical. Showing those results, the video clips of the individual trying to gain access to the garage, different units, flashing her badge as she moved through — those were powerful learning moments. We can have the best policies and procedures in place, but if we’re not actively testing them, we won’t be successful. Consistent testing and validation of your cybersecurity program is critical. We’ve focused heavily on deepfake and phishing awareness — phishing emails are getting more advanced and harder to detect because threat actors are using AI to create them. That’s why I’ve always said: we have to fight AI with AI. We focus on EDR, SIEM, SOC, identity management, two-factor authentication, and micro-segmentation. And if we do not invest in our human firewall, we will not have a successful outcome, because your end users are either letting threat actors in or keeping them out. Building that human firewall is one of the most important things we do.

Ritu: Thank you, Michael — that was really comprehensive. Rohit, would you like to ask a question?

Rohit: Thank you, Ritu. Michael, I was thinking about what you said in relation to the use of generative AI tools like ChatGPT by employees. We know everyone is embracing these tools in their personal lives and bringing that into the business environment, where they might inadvertently expose proprietary information. What are your thoughts on how generative AI is being embraced organization-wide, and what AI governance initiatives do you have in place?

Michael: Generative AI sites like Claude and ChatGPT are currently blocked within our entire environment. There have been breach issues with ChatGPT where individuals put in proprietary information — if an end user doesn’t configure the session correctly in OpenAI, the algorithm absorbs that information, and theoretically it could be leaked, creating a massive breach. So we’ve blocked all those tools. Right now we’re investing more in Microsoft Copilot, which is being rolled out across the organization, integrated into our Office 365 environment and customized to how we operate. Free-range use of general generative AI tools could create a major security issue. Copilot is the approved application with all the appropriate security processes in place — all other generative AI tools are completely blacklisted.

Rohit: So people have access to Microsoft Copilot for regular use?

Michael: That is correct — it’s a designated, approved application with all the security processes we’ve developed in place. All other generative AI tools remain blacklisted.

Rohit: How are you seeing adoption of Microsoft Copilot across departments?

Michael: There are a lot of great use cases. Departments are using it for finance validation, working directly in spreadsheets, developing reports and breaking them down into more readable formats, generating ideas, enhancing policies and procedures, and validating information. Departments are customizing how it fits their specific workflows, how to accelerate their work, and how to make things more efficient. It has been a really good tool for the organization.

Rohit: One more question — what are some of the innovation initiatives or processes you follow, and how do you foster innovation within the organization?

Michael: One of the biggest initiatives we’re focused on right now is the RHTP program — the Rural Hospital Transformational Program. It’s about $50 billion over five years, each state applies separately, and the State of Colorado received approximately $160 million. Critical access organizations are going to apply for this, and it’s really going to transform how rural health delivers healthcare services. This is where you’ll see a lot more innovation coming out of rural America — healthcare has gotten behind the curve on digital transformation, and in rural America the pace has been even slower. These funds will allow organizations to move in the right direction if they strategically align their projects with the application process. For us, a big driver is focusing more on remote patient monitoring and in-home hospital initiatives. We have many remote patients who have difficulty coming in and tend to wait until things get worse before seeing a provider, rather than having their health consistently monitored. As I said from the start, your zip code should never determine your healthcare outcomes — but in rural America, people statistically live shorter lives than those in metro areas because it’s very hard to access the same level of services. Our model has been: how do we focus, strategize, innovate, and bring technology that actually matters to the organization? We’re a CMS five-star hospital — one of only two rural facilities in Colorado to receive that designation. We’ve been a beta site and an innovator, doing more with less. We’re incorporating AI to help providers select appropriate CPT codes based on their completed documentation — before AI, providers were essentially acting as billers and coders, which took them away from being doctors. Reducing those barriers, automating the right things, and getting to a place where your zip code truly does not determine your outcomes is the philosophy and the direction we are moving in. The acceleration of digital transformation is definitely here.

Ritu: Totally agree. One final question: CIOs are now expected to oversee not just traditional IT, but AI, cybersecurity, cloud strategy, and digital experience. How do you see the role of the healthcare CIO evolving over the next couple of years, and which leadership capabilities will become more important than technical expertise?

Michael: I am a technology executive, but we have to be business leaders first. We have to understand the organization as a whole — not just from a technology standpoint, but as a complete operational entity. The CIO is increasingly focused on creating automation that enhances revenue cycle initiatives, and on moving away from the old stereotype of IT as purely a cost center versus a strategic revenue contributor to the organization. The evolution is about focusing on better patient care outcomes, recruitment, reducing burnout, business continuity, and operational excellence. Everything we do now has a technology piece associated with it. But the CIOs who will really make a difference are those using technology for better outcomes — technology that truly matters, technology with a patient-centric initiative attached. Those are the game changers and the innovators moving forward as the CIO role evolves.

Ritu: Thank you for that answer, Michael. That brings us to the end of the podcast. Thank you so much for being our guest today.

Michael: Thank you very much. You both do an amazing job.

Rohit: Thank you, Michael.

—

Subscribe to our podcast series at www.thebigunlock.com and write us at info@thebigunlock.com    

Disclaimer: This Q&A has been derived from the podcast transcript and has been edited for readability and clarity.

Rohit: Hi, Alex. I’m Rohit Mahajan, co-host of The Big Unlock Podcast and CEO and managing partner at Damo Consulting. It is great to have you as our guest, especially at this fantastic event, AI4. I would love you to give your introduction to our listeners.

Alex: Thank you very much for hosting me — I’m a big fan of the podcast and when we connected I immediately decided we had to do this. I’m Alex Zhavoronkov, founder and co-CEO of Insilico Medicine. We are a generative AI company that is clinical stage, meaning we actually have real drugs in clinical trials. We are publicly traded — we went public in Hong Kong last year under the ticker symbol 3696, so I need to be careful about forward-looking statements. My background is in computer science, and I started my career in GPU computing. I made some money in my early twenties and decided to spend the rest of my life on aging research, because I believe that’s the most important problem everyone has — it is the major cause and source of suffering in the world. I did my graduate work at Johns Hopkins, was a professor here and there, and in 2014 came back to my roots in GPU computing and started Insilico at the NVIDIA GTC conference. We had a presentation called “Can NVIDIA Solve Aging?” Fast-forward twelve years, we’ve managed to get a few drugs into human clinical trials, and I’ve raised a considerable amount of funding — though it wasn’t easy in the beginning. There were times when I had to sell everything and put it into the company. When you’re rich you invest everything, then you’re poor, then you get it back — and then you don’t want to be rich, so you invest even more. It’s an interesting journey. I’m still far from solving aging, but we’ve made some good progress.

Rohit: That is a very interesting area. Everyone talks about the silver tsunami — 10,000 Americans aging every day, going into senior care or rehabilitation facilities. When you say you’re working on longevity, there is health span and there is lifespan. Tell us how you think about that, which drugs are in the pipeline, and how you chose them.

Alex: I actually focus on lifespan — for me that is more important than health span, and I’ll explain why. Health span is a very vogue definition of longevity, basically used by people who don’t want to push the boundaries of actual lifespan. I don’t know of any drug that can give you ten years of life without also giving you additional years of healthy life. If your goal is to push that boundary, you’re also going to increase health span no matter what — the question is just the size of the delta. If that delta is six months of suffering, that’s not something we’d want to focus our entire company on. That said, we still develop drugs that may help in certain conditions, because very often diseases are diagnosed too late. I would still give six months to a mother of two suffering from terminal cancer to have more time with her children — especially if it doesn’t cost much, and it has to be her choice. But we need to provide people with the freedom to live even six months longer. People who say they don’t want lifespan, only health span, are actually doing damage to the field, because we want lifespan too. And if you want to be truly ambitious, think about peak span — how long you spend within ten percent of your maximum performance. All of those spans matter and there shouldn’t be a debate about which one to extend. Life is a fundamental human right. If we can give one quality-adjusted life year to everybody on the planet, that is 8.3 billion life years — 110 million lifetimes. A couple of years would be 220 million lifetimes, which is more than were lost in all the wars fought in the history of humanity. The way to get there is to work within the traditional pharmaceutical drug discovery and development paradigm — going after a disease, getting a drug approved — but if the root of that drug was actually aging research, you’re developing a longevity therapeutic with AI. The AI tells you this is an anti-aging drug but it should work on this cancer, so you approve it for the cancer. The patient gets additional months or years. And then once the drug is approved and safety and efficacy are established, you can experiment with it, gather real-world data, and see if it also improves aging biomarkers — potentially beneficial for others at a different dose or mechanism of administration. That’s the idea.

Rohit: I heard you say longevity and cancer. Are most of the drugs in your pipeline targeting different cancers, or something else as well? What’s the breadth and depth of what you’re going after?

Alex: We are very different in the context of AI drug discovery. Most companies go after cancer or neuroscience. We are the MMA fighter — we decided to go very broad spectrum. Our core ideology is that we go after aging. Many of the protein targets that drive disease are implicated in both aging and cancer at the same time, just used differently. For aging, for example, you want to eliminate senescent cells — cells that have stopped functioning and are just sitting there excreting toxins into the microenvironment. Some cancer drugs do exactly that. About half of my pipeline is related to cancer and half to chronic disease. My lead program, currently in phase three, targets idiopathic pulmonary fibrosis — a chronic lung disease with no good treatment at this point, where everything available just slows the decline rather than reversing it. In our phase two trial we demonstrated very promising reversal of the loss of forced vital capacity, the essential measure of lung function. We also go after IBD, inflammatory bowel disease, neuroinflammation including Parkinson’s and potentially Alzheimer’s, ocular diseases including dry AMD and uveitis, and the most exciting recent breakthrough is in pain. Pain is very difficult — identifying a new mechanism that isn’t opioid-based or an ion channel or anti-inflammatory is extremely hard. Most pharmaceutical companies historically started as painkiller companies, and that’s how we got heroin, morphine, and fentanyl. We identified a new target originally purposed for aging, and through sheer scale of experimentation — while testing a pain drug on animals, we had the ability to put more compounds in the assay. In one experiment we rescued an animal in terrible pain with something that might not have worked, and it worked better than morphine and epidural. We tested it against every standard of care, even against NAV 1.8 orally, and it performed significantly better. We took it all the way to a development candidate — one step before human clinical trials — and it went great. We now have 32 development candidates, eight phase ones, three phase twos, and one phase three. The pain drug alone is extremely exciting, because imagine if at high dose it kills pain and at low dose it addresses aging.

Rohit: That’s a great combination. We’re at an AI conference, so tell us more about how you’ve combined the different AI approaches. AI was once predictive analytics, now it’s generative — and you’ve been finding targets and small molecules for drug discovery for a long time. What’s novel about how you’re approaching AI?

Alex: Insilico originally started as a deep learning company and closely followed DeepMind — when we hired people, we tried to do hackathons where we’d take one of their papers and hire people who could perform similarly. We started as a biology company with an AI biologist focused mostly on understanding how you live from birth to death using different biological data types. Those were originally predictive systems. In 2016 we started publishing on generative adversarial networks — early days of generative AI — and we also purposed those algorithms for chemistry for the first time. My first paper on GANs for multi-parameter optimization and molecular generation came out in 2016: instead of searching for a needle in a haystack, you generate perfect needles — molecules with desired properties. In 2017 we published papers showing experimental validation of the technology: we actually synthesized and tested the molecules. In 2018 we published in Nature Biotechnology on generative tensorial reinforcement learning — right before ChatGPT — showing we could synthesize and test molecules all the way into mice in 46 days. That was a big deal. From the early 2020s we started working on transformers and diffusion models, but essentially built a Lego system of different models that can do biology and chemistry. Some do generation, some do synthetic data generation — because in biology you often don’t have much public data for specific problems, so you can actually generate high-quality data using generative approaches. Some models do predictive analytics, some predict clinical trial outcomes. We orchestrate all of them using frontier models. We now have over 1,200 tasks in drug discovery — think of 1,200 experts you can clone and spawn into many different agents to achieve the grand task of reasoning across an entire program from target identification to approval, working backwards. Essentially: from prompt to drug.

Rohit: “Prompt to drug” — that’s a great phrase. I’ve been following this space and I think our listeners would also ask: how do you position yourself relative to Flagship Pioneering, Moderna, or other companies doing pioneering work in this space?

Alex: Flagship Pioneering is a great platform for company creation. What they do is essentially have capital, a pool of experts, and they identify trends — for each trend they build a company, offer it to investors, raise funding, and utilize the same people incubating many companies. In generative AI they have Memong Mini, incubated by the same team for different trends and investor bases. I’m not entirely sure how to establish the success of those companies. One called Generate Biomedicines is listed publicly and has a drug in phase three, though it would be nice to see peer-reviewed publication of how much the generative approach actually contributed. Others I’m less certain about, because when you create a company for a purpose it’s very difficult to get people to work together toward a specific goal. Companies that tend to be more successful in our field are those where people came together, worked on a problem, solved it, and only then raised money to scale. Forced invention is very difficult — genuine invention works better. That said, Flagship is a great platform, and one win like Moderna pays for the entire party. Moderna is actually a Flagship company, which is why they’re famous. And if it hadn’t been for COVID, it would be an open question where Moderna would be now. But the companies I’m actually more concerned about are Anthropic, OpenAI, Google, and many Chinese players like Tencent and Alibaba. They are developing foundation models that can reason really well in the context of biology. Some of the more primitive tasks — like target discovery — are already completely demonetized. More complex tasks are being demonetized as we speak. All of those companies are also putting resources into biology. My hope is they don’t make the mistake of buying low-quality companies just for the data or for kudos — those are computer scientists who don’t know what works in biology. But once they get proficient in our field, they will be real competitors. What I’ve started doing is developing tools that actually accelerate this transition — tools like MMI Gym that help frontier models train on what we do. If in some tasks they are advancing and can be better than me, I’d rather help them do that, because they can help us back. I’m going to be always at the frontier with a new algorithm or approach. In the future, the real moat in my field are the drugs — they are like diamonds, they’re forever. AI comes and goes every six months. The future moats will be capital — which you can convert into energy or compute — infrastructure including labs, robotics, and networks of contract research organizations, and third: good intent, good ideas, and being in the right place at the right time. We have more ideas than we have capital or infrastructure, but with increased intelligence we can more rapidly convert ideas into real products that save lives.

Rohit: That’s a great insight. What are some of your biggest challenges at the point in the journey where you are now?

Alex: One fundamental challenge that everybody faces is the complexity of biology. You need to try a lot of things to see what works, even with perfect AI. I can now go from prompt to drug for a given target — if it’s low or moderate novelty, I’ll win. But if I’m going into truly new terrain, there’s a very good chance I’ll fail. You need a very sustainable business model to allow yourself to fail. The real great challenge I see is geopolitics. Right now it’s absolute nonsense. I don’t understand it and don’t want to understand it, but I have to. The US is fighting with China, countries have become very nationalistic. To do really good work in our field, you need international reach. If you want to synthesize molecules at scale, there are only two places you can do that — India and China. You can’t do it in the US; the infrastructure simply doesn’t exist, and it’s also a low-value task. Some animal experiments, like primate studies, can’t be done at scale in India either. Most of the hardcore competition is in China, so if you want to compete you actually want to compete there. The US is making it difficult for American companies to do that, and there are regulatory complications everywhere. My job isn’t to pick any side — I don’t care where you live, as long as you can live longer. If a mother of two is dying somewhere in Africa or in China, you need to help her. We’ve had to establish infrastructure that allows us to be global — we’re in Montreal, Abu Dhabi, Hong Kong, Taipei, Shanghai, and Yixing. Right now it’s just very difficult to operate globally, and that difficulty is a real challenge for what we’re trying to do.

Rohit: Any upcoming announcements or plans you’d like to share, including any plans to list in the US?

Alex: We would of course love to explore additional capital markets and are constantly on the lookout, timing the markets carefully. The US biotech industry is going back up but it’s still in a winter — it hasn’t fully processed the excess from companies that listed and raised a lot of capital in the early pandemic days, didn’t deliver, and lost investor trust. The AI hype also needs to settle somewhat, because right now people are chasing trillion-dollar companies and forgetting about smaller biotech, even though it’s very important. What we have on the horizon are massive scientific breakthroughs we’re constantly working on — but as a publicly traded company I can’t talk about them specifically, so watch for peer-reviewed publications. We usually don’t make big claims until we publish. What excites me most are the clinical trials. Once you’re in the clinic, you’re worried all the time — with many programs running, you must fail statistically at some point. So far we haven’t, but given our current rate of success I think in many cases we should succeed, and when we do it pays for everything. On the AI front, our most important initiative is benchmarking — we just released a set of benchmarks where we can test frontier models and specialist models across over 1,000 drug discovery tasks. Many of them perform poorly; some are reasonable. The large foundation model developers don’t even know drug discovery yet. I’m very happy to see that Anthropic is actually going into their own drug discovery — you need to discover a drug to know how to discover a drug; it’s the chicken and the egg. The most exciting thing for me remains aging research. I don’t think there are greater enemies that humans have other than aging — it will kill you with one hundred percent certainty and takes everything away. There’s a good chance we can give everyone an additional ten or even twenty years. We have drugs in development that hopefully will get us there. That’s what will consume a large part of my life, and I’m willing to fight for it.

Rohit: That’s beautiful — pushing the envelope on longevity. Thank you so much, Alex. Really appreciate you being our guest on The Big Unlock Podcast.

Alex: Great to be on the podcast. Let’s unlock longevity.

Rohit: Yes. Thank you.

—

Subscribe to our podcast series at www.thebigunlock.com and write us at info@thebigunlock.com    

Disclaimer: This Q&A has been derived from the podcast transcript and has been edited for readability and clarity.

Ritu: Hello, listeners. Welcome to Season Seven of The Big Unlock Podcast. My name is Ritu Oberoi, and I’m managing partner at Damo Consulting and your host today along with Rohit. Today we are really excited to welcome Zaw Thet to our podcast. Zaw is the co-founder and CEO of Exer AI, a digital health company using computer vision and AI to assess, coach, and improve human movement. Zaw brings an unusual combination of experience as a serial entrepreneur, investor, and operator, having previously been a founding partner at Signia Venture and CEO and co-founder of ForInfo. His current work with Exer sits at the intersection of musculoskeletal health, physical therapy, sports performance, and AI-powered care delivery. Welcome, Zaw — really excited to have you here today.

Zaw: Thank you so much for having me. When you say all of those things out loud, it makes me blush that I’ve been doing this now for twenty-five to twenty-seven-plus years. It’s a pleasure to be here.

Rohit: Thank you, Ritu. I’m Rohit Mahajan, CEO and co-host of the Big Unlock Podcast. Welcome, Zaw, to the podcast. We really appreciate the friendship and are looking forward to an engaging discussion.

Ritu: Great. Zaw, tell us about Exer, which is built on the idea that movement itself can become measurable through ordinary cameras — without wearables or any extra hardware. What does that unlock in healthcare that we previously just couldn’t do when movement assessment depended mainly on in-person observation or wearing devices?

Zaw: Thank you again for having me. To explain why human motion matters, I need to rewind to about eight years ago when we decided to start the company. As an investor looking across multiple digital health disciplines and companies, I was considering the different potential applications for AI — this was before ChatGPT, Claude, or any of the current LLMs and large models were out there. What we were really looking at was discrete sets of data, because I believe all AI is based on the moat around the data it has been trained on. What became very evident as we talked to people across multiple disciplines was that there was a ton of data in the chart. All the companies looking at EKGs or blood panels were essentially training on the same types of data. There was a lot of early AI starting to emerge around anything related to the back office of health system operations — revenue cycle management, administration, prior authorization. There was also a lot around digital scribing, whether ambient scribes listening from the room or simple dictation tools like Nuance. But what became clear was that in all the imaging being done — X-rays, CT scans, MRIs — actually seeing dynamic human motion was something that really didn’t exist. It existed in gait labs with multi-million-dollar cameras and sensor rigs, but deploying that on the edge or in clinical settings for a mass amount of patients wasn’t possible. What we heard over and over again was that even though the data didn’t exist yet and people didn’t fully know what it would mean, it would be really helpful to capture human motion in a clinical environment, at a patient’s home, or over telehealth, and understand what that means for different diseases. That’s the basis on which Exer was formed — with the understanding that a lot of research would be needed to validate all the different algorithms for assessments and phenotyping across different disease types. That’s how we got started, and it has obviously evolved a lot since then.

Ritu: You’re not just building another healthcare app — you actually want to turn the camera into a scalable clinical and coaching tool for movement health. How does real-time AI coaching change patient engagement and outcomes compared with traditional home exercise programs or simply telling patients what to do?

Zaw: A lot of people in the early days mistook us for a PT company because many of the assessments are around fundamental human motions — a hinge, a squat, a pull, a gait walk. But no one had really thought about how to apply that to clinical settings as a whole. When we first started, the feedback we got from people was: if you want to make this work inside of very large health systems — and we’re now inside twenty-five of the top fifty health systems in the US — how do you deliver a platform, not a single-point solution? A platform meaning it doesn’t just apply to one department but applies to orthopedics, neurology and neurosurgery, primary care, PM&R, pain and spine, and so on. The second was: how do you actually deliver on the three legs of the stool? How do you deliver demonstrably better patient outcomes, better clinical efficiency — actually taking time that would normally be spent by a clinician, nurse, PA, or MA and making that more effective so they can scale across more patients — and potentially increase new revenue streams for a hospital where they may not currently be capturing that care in-house? Those were the two primary drivers. The other requirements were par for the course: regulatory approval as a Class 2 medical device, SOC 2 and HIPAA compliance, and clinical validation through peer-reviewed publications. The way we approached it was to build two different parts of AI. One is the neural net that allows us to see the human body without a gait lab — just using a phone or iPad. The other is delivery across multiple points of care, primarily in-clinic in outpatient settings, with telehealth and at-home options as well. Then we built a research arm across multiple disciplines. I’ll give you a quick example. Total knee surgery is probably the most common orthopedic surgery in the US, with over a million done every year. The current standard of care: you visit your orthopedic surgeon, they assess your knee, determine you’re a candidate, schedule surgery for three months out, and send you for a long leg X-ray — a static image taken while you’re standing straight with your kneecaps together. That image is used to measure the angle from hip to knee to ankle, but it doesn’t capture how you actually walk — whether you have a bit of varus or valgus. None of that is captured by an X-ray today, even though research has shown it’s not a great predictor of surgical outcomes. Instead, imagine having an iPad in the clinic room. When the patient agrees to total knee surgery, a medical assistant can assess their gait in under twenty seconds and understand their circumduction patterns and how the knee loads at maximum weight-bearing — all of that populated directly into Epic or whatever the EMR is before the doctor even comes into the room. That assessment is repeated pre-surgery, checked immediately post-operatively to confirm alignment with the angles the surgeon intended, and tracked again three months later as ligaments relax. That’s just one example in orthopedics of how this is already being deployed today.

Ritu: Wow, that’s an amazing example. A static picture really can’t tell you that much compared to seeing movement on an iPad. How does the telehealth part work — if the patient is at home and you’re doing a remote appointment, are you just recording them on a normal browser and still getting all the information?

Zaw: Telehealth and at-home work typically over a web browser — the easiest approach. It’s done inside any modern browser, Chrome, Safari, and so on, using the same basic tools. It’s a little different because the clinician is on the telehealth call with the patient, and it can be a bit harder for the patient to set up in their home environment. But especially for anything upper body, we can capture it well. One of the big telehealth use cases is in neurology. We can track 22 discrete points on the hand, which makes this very powerful for orthopedic and post-operative visits — especially for destination hospitals like Mayo Clinic where patients have flown in and don’t want to travel back just for a follow-up. We can track where they are on their recovery curve: whether the hand is achieving full range of motion is very difficult to assess even on physical exam using a goniometer for each finger joint. Our platform allows that to be done instantaneously. Another great example is tremors on the neurology side. Because we can see the full body, we can track how a medication is affecting a tremor as a patient is titrating different drugs, as well as other balance and ataxia measures. That’s how the telehealth piece works, and there’s also an at-home option that works over a browser or as an app on the patient’s phone.

Rohit: How did you get started in this space, and are there any prior ventures in healthcare you’d like to share with the audience?

Zaw: I don’t come from a traditional healthcare background — I come from a software startup and tech background, though I do come from a family of physicians; both of my parents are physicians. Like every good first-generation immigrant, I was supposed to become a doctor, an engineer, or maybe a lawyer as a third option — and I chose a different path. When I went out to Stanford in the late ’90s during the first dot-com boom, in the early days of eBay and before Google even existed, I saw an opportunity. I’d always been interested in technology, and what I found was my true calling: solving big problems in society using technology, where you can scale the impact to millions of people. When I first considered medicine, I loved the mission and purpose and the ability to help people — that has always driven me. But I wanted to deliver that at scale. The problem with a one-to-one ratio between time spent with someone and help given is that you can only help that one person. My superpower, I guess, was the ability to build big platforms where instead of helping one person I could hopefully help a thousand. That’s the big vision we had when we got into healthcare — I saw a lot of inefficiencies and maybe I was a little too naive, thinking this is a big problem that needs fixing. I didn’t realize how complicated it was eight years ago. I joke that I still don’t have a full healthcare license — more like a green card — and I’m learning every single day. But what I found was that the ability to deliver precision medicine at scale is especially critical given the severe shortages of doctors, nurses, and PAs alongside an ever-aging population that only increases demand on the healthcare system. Those two things are fundamentally broken and won’t fix each other. Yet somehow healthcare still works — survival rates for cancer and cardiac events in the US keep going up. The medicine is getting better and we are helping people, but it’s a broken system. My hope was that Exer could be one component of fixing it — allowing clinicians and providers to deliver better medicine at scale, working with them and not against them. That was a core tenet from early on.

Ritu: Exer has so many applications across physical therapy, senior care, orthopedics, and sports. So far, which market has been the strongest, and which do you think becomes the biggest long-term opportunity?

Zaw: People always asked us this, especially in the early days. When we had our first demo running in 2018, no one had ever seen anything like it, and everyone wanted a better golf swing — or tennis serve, baseball or lacrosse. We looked at all of those early on and said we could build a consumer company doing very specific sports applications, or we could start where the science is really hard. If we can prove we can do this for the top doctors in the US, and that’s where the research originates, then we have the opportunity at a later point to go anywhere we want. Think of Amazon starting by just selling books — that’s sort of the first inning we think we’re still in today. We focused on healthcare and on delivering a platform to health systems and providers rather than selling directly to consumers. Maybe that was the harder route, but it also meant we didn’t have to spend enormous amounts on marketing or raise hundreds of millions of dollars, and it let us keep the team lean — because healthcare just takes a long time regardless of how much money you throw at it. Our secret advantage is that we now have four-plus years and hundreds of thousands of patients who have run through IRBs and clinical studies, giving us a very proprietary data set that no one else has — where clinicians and researchers are using Exer to record not the video but just the body points as they move, tied back to all the clinical data from the chart. We understand what a gait shuffle looks like: does it mean arthritis, stenosis, or Parkinson’s? To do all of that and also do golf would have been very difficult. We’ve stayed in our lane, and we’re going to be there for a while because there’s still a lot to do in healthcare before we get to the, quote-unquote, fun stuff.

Ritu: This leads nicely into the next question. We’re talking about continuous functional monitoring, and wearables are having a big moment. We’ve talked to clinicians and doctors about how outdated the annual physical is. How do you see this tying into more continuous monitoring of patients and catching problems earlier?

Zaw: I think it ties in really well. We’ve been fans of working and partnering with companies that are more device-specific — consumer companies like Whoop or Oura Ring, medical-grade companies like SensorBio, and data coming off a watch like a Garmin or Apple Watch. There is a lot of room in this space for continuous monitoring. It’s already happening at a certain level for patients at home managing diabetes, for example. The challenge on the continuous monitoring side is filtering through all the noise — there are a lot of signals potentially coming from these devices, and knowing what matters most is the hard problem. That will get solved, but probably not by a company like us at Exer. Where Exer comes in is that a big component of that annual physical is a movement screen — especially for a senior population 65 and over, where you’re looking for issues and particularly fall risk, which is related to gait and balance. Falls are the number one cause of death and the number one cost to our healthcare system for people over 65. If we can prevent that, if we can get ahead of that curve — instead of checking someone every 12 months, doing a movement screen once a month — all of their wearable data is being pumped in at the same time. Low blood pressure that may cause fainting comes from the wearable; arthritis in the right knee that could cause a fall comes from us. That is definitely the future of medicine, and it will get there over time.

Rohit: The AI landscape is changing so fast. Could you give us some thoughts on how to keep up, and how you’re thinking about incorporating new developments into your offering for clients?

Zaw: It’s moving at a phenomenal pace. When we first started, we were really one of the first true AI companies in healthcare. Because of the research and time it takes to be clinical-grade and deployed inside health systems, we’ve had to be very patient. Now you’re seeing algorithms coming out left and right — some just based on large LLM models, some based on real data. You have companies like Open Evidence helping physicians find information faster. It is a brave new world and very hard to keep track of. I get the email digests from Silicon Valley and look at the top developments. For patients, most of these things are happening without them seeing it. There are starting to be some patient-facing applications that are genuinely interesting — like scribing your own doctor visits, then feeding that recording into an LLM to better understand what was discussed. It’s probably better than playing Google Doctor. But for the most part, most of the AI patients see in healthcare is behind the scenes. Exer is actually one of the exceptions — when you come in for a visit, you’re being assessed, and there’s a playback feature that’s one of the most-used things after an assessment: the clinician pulls it up, goes through it with the patient. It looks like a skeleton moving, so the patient can see what they’re doing without any actual video. Some patients even like that — it removes concerns about how their hair looked or what they were wearing. As a business owner using these tools all day, they’re phenomenal for making a business punch above its weight. A ton of our customer support is now handled by AI agents. Our developers are at least two to three times more productive because a lot of junior developer tasks — QA, code checks — are all being done by AI. There are countless ways to incorporate AI into your business, and it will only keep evolving. We tend to find one we like, stick with it, and try not to switch as often as possible.

Ritu: We’re almost at the end, Zaw. Any closing thoughts or crystal ball predictions for the next year?

Zaw: I’m a big optimist — I wouldn’t be doing this if I didn’t believe we can change the world for the better. That said, I do think there’s some potential backlash coming that we need to be aware of, especially in healthcare around the use of AI. I’m already hearing early pushback from systems, particularly around generative AI where the AI is generating content and having a patient interaction without any human in the loop at all. The capability is really good, but a small percentage of the time it halluccinates or sends inappropriate information, and we’re starting to see the first lawsuits around that. I don’t think that means it will go away, but for anything without a human in the loop there’s going to be a period of pushback about whether this can scale directly to consumers and who bears liability when it sends something incorrect. A lot of that will get sorted out in the next couple of years. At Exer, thankfully, there’s always a human in the loop because we provide support to clinicians, not replace them. In the end, I think we’re obviously going to end up with a much better and more efficient system. That’s my crystal ball.

Ritu: You raised very good points. The ability of generative AI to hallucinate leads to real ethical questions — the moment you remove the human from the loop and cede that control, it needs to be fail-proof, which it isn’t yet.

Zaw: We’re going to have agents watching agents watching agents, and then maybe a human to check at the end. Who will guard the guardians — that’s right. Well, thank you both for having me. I really appreciate it.

Ritu: Thank you so much, Zaw. Thank you for being on our podcast.

—

Subscribe to our podcast series at www.thebigunlock.com and write us at info@thebigunlock.com    

Disclaimer: This Q&A has been derived from the podcast transcript and has been edited for readability and clarity.

Ritu: Hello, listeners. Welcome to Season Seven of the Big Unlock Podcast. My name is Ritu Oberoi, and I’m managing partner at Damo Consulting and your host today. We are looking forward to a conversation with Dr. Angela Shippy, and we warmly welcome her to the podcast. Dr. Shippy has been a director and senior physician executive and clinical innovation lead at Amazon Web Services for over five years. She leads executive engagement for healthcare and life sciences. Her work focuses on bridging operations and technology to enable and accelerate transformation for providers, payers, pharma, biotech, health tech, and device organizations. We’re really looking forward to a very engaging conversation with Dr. Shippy, so welcome to the podcast once again.

Angela: Thank you, Ritu. It’s great to be here, and thank you for having me.

Ritu: Dr. Shippy, would you like to add anything to that introduction?

Angela: I think the one thing I would add is that people often ask how I ended up at Amazon Web Services. Truly, it was because I have been an end user of technology my entire career. I started my residency a few days early because we were all mandated to use a very early version of CPOE — computerized physician order entry. All along the way I’ve continued to use technology, whether it was being a physician champion for PalmPilot, electronic documentation, or putting together templates with the informatics team. It’s always been part of my career. And like so many of my colleagues, the COVID pandemic was really an opportunity to bring data together. For the first time ever in my previous organization, we brought together clinical and operational data in a dashboard that we were using on a regular basis. Seeing in real time how having all the right data in the right place allowed us to treat a patient population that was in need — and know exactly what they needed — really spurred me to join this team and help my colleagues across the country, and across the world, be able to do this.

Ritu: Thank you for sharing that. We would love to hear your origin story — how did you decide to become a physician?

Angela: It’s a funny story if you think about a nine-year-old girl who had a teacher who told her she was really good at science. My fourth grade teacher said, “You’re really smart and you’re really good at science.” I translated that to mean that if you’re really good at science and you’re smart, you should be a doctor. And then I decided that if I was going to be a doctor, I should be a neurosurgeon because the brain is the hardest thing and the most important part of the body. For many years through elementary school, I was focused on becoming a doctor and a neurosurgeon — I used the term brain surgeon at the time. In high school and college I shifted to thinking I would love to be an OB-GYN and deliver babies. When I got to medical school, it was my very first rotation, and I loved it — a six-week rotation starting in labor and delivery, so much fun. Then I got to the surgical part: bladder repairs, hysterectomies, and oncologic patients, and I realized that wasn’t the part I was excited about. My next rotation was internal medicine, and I loved it. Internal medicine is the gateway to so many specialties, and I ended up as an internal medicine physician, practicing as a hospitalist for many years.

Ritu: Thank you for sharing that story, and a big shout-out to any teachers out there who are listening — those words set you on this path. So Dr. Shippy, you bring together the perspective of a practicing physician, a former health system executive, and a technology leader to help organizations think about what’s possible in this new era of cloud computing and generative AI. In your experience working with so many health systems, what separates the organizations that are truly transforming with AI from those that are just trying pilots and deploying isolated use cases?

Angela: Right now so many of our customers are thinking about their biggest pain points — what their clinical operations look like, what throughput looks like, whether they have enough nurses, enough allied health professionals, and what they need across physician specialties. They’re looking at the entire picture, and that’s leading them to ask: what do I want to do with generative AI and agentic AI to take things to the next step? The organizations doing the best work are those asking two things. First, which particular problem am I working backwards from to solve? Second, what am I doing to scale beyond just that initial pilot? I’ll give you an example. UCSD is using one of our purpose-built solutions for healthcare — Amazon Connect Health — allowing them to engage with patients in a meaningfully different way. Being able to verify the patient when they first call the contact center, direct the call to the right part of the organization, use all the data they have related to that patient to direct them appropriately or provide a quick, accurate answer. For them, this has meant 3.2 million contacts with patients, a reduction in hours that call center staff spend on those calls, and a thirty to sixty percent reduction in abandoned calls depending on the department. That translates to patients getting to preventative care, managed care, prescriptions — whatever they need — faster and more accurately, which produces a better patient experience. And we know that patients who have a better experience have better health outcomes. When you see that level of engagement with patients and their ability to access the parts of the healthcare continuum they need, you also know those are patients who continue with that same health system — which addresses the operational efficiency goals as well. The systems really advancing beyond pilots are those focused on what matters most.

Ritu: Thank you for telling us about Amazon Connect Health and how you’re using autonomous agents for patient navigation, appointments, rescheduling, and ambient capabilities. That leads into the next question: AWS has invested heavily in helping organizations build scalable data foundations before deploying AI. But so many executives want to jump straight to the next shiny new AI application. How do you help healthcare leaders resist that temptation and understand they first need to build foundational capabilities?

Angela: One of the things I really enjoy about my work is the opportunity to talk to colleagues across the country about what they’re doing. In those discussions, we talk about the end goal — but keeping in mind that to meet those goals, they first have to think about their data. Do they have clean, available data as their foundation? We know that right now 97% of the data in healthcare does not get used, largely because it’s unstructured and from so many disparate sources. Part of that initial conversation is about how to bring all that together — how to take that disparate data and turn it into real information that becomes insights. Then being able to use agentic AI to take that data and information and turn it into actionable insights, whether that’s for the patient experience, the clinician’s experience, or for research. For a payer organization it can help with member and population management. For a pharma company it can help get to drug discovery faster. That’s what we talk about — the foundation of clean data in the format that allows you to use these agentic tools.

Ritu: You mentioned that you started with technology training even before your residency. Today’s CIOs, CMIOs, and digital leaders need to understand cloud architecture, AI, and cybersecurity — all changing at a rapid pace, sometimes weekly. What is your advice to CIOs and CMIOs on how they stay abreast and relevant when the pace of change is so rapid?

Angela: Our customers share a curiosity about what’s happening from a technology standpoint — they’re standing side by side with us wanting to learn. We have thousands of customers running workloads on AWS, and we’ve earned their trust over time, so they see us as a source of information. We feel a commitment to help them continue to learn in this rapidly changing environment. We do that a few ways. One is through summits — I’m at the DC Summit right now, which gives customers the opportunity to come and learn, hear from their peers, and see what others are doing. We also have training and certification available for leaders and their teams to use — much of it free. We have a fellowship program where up-and-coming leaders from health systems can come through and learn more about technology and, as they advance in their careers, how they’ll think about working with technology and specifically with AWS. We’re also dedicated to upskilling, because every worker within healthcare and life sciences will see their day-to-day job change because of the tools that are available. We want to make sure they have an understanding of that. The reason this is so important for CIOs and CTOs is that we know there’s significant administrative burden across the workforce they’re responsible for, and these tools can help decrease that. When you’re at a summit like this, the energy is remarkable — especially across public-sector customers who are mission-driven, highly regulated, and are really looking to AWS to be the partner that helps them in secure environments, with managed AI layers that allow them to deploy quickly without having to dig into every single aspect of the technology, and with purpose-built services that have guardrails in place for another level of security. All of that is what we’re doing to ensure CIOs and CTOs have knowledgeable teams, that we’re supporting those teams, and that we’re supporting their organizations and goals.

Ritu: So you’re saying this technology can really be an enabler and help them do their job better.

Angela: Absolutely — it’s an enabler and an accelerator. It allows them to start with a proof of concept and then truly scale it across the entire organization. Right now we’re working side by side with customers to help them meet these goals, and we want them to look to AWS and know we have solutions ready for them. I’ll give you another example — Amazon Bio Discovery. This is a new technology that allows our pharma customers to move so much faster. The time it takes to research, take all the data from that research, and get to the discovery of a new drug takes years. With Amazon Bio Discovery, being able to put all the components together and help them get to the answers and insights they need so much faster is going to decrease the time it takes to develop the next drug. That might mean the next drug we need in the oncology space is developed much faster — more lives saved. It’s exciting to think about, and we really look forward to continuing to take this drug discovery work and ask: what can we do next, what can we learn, and what can we iterate and innovate on in this space?

Ritu: Healthcare leaders often talk about moving from data to insights, but increasingly the conversation is shifting toward moving from insights to autonomous action. As cloud platforms evolve to support AI agents and intelligent workflows like Amazon Connect, how do you envision the role of health systems changing over the next few years? Will we still be thinking in terms of software applications, or will patients see seamless, almost invisible technology that just helps them — without passwords, thousands of clicks, and getting stuck at every point?

Angela: I think right now we will see an evolution — technology will become easier to access, and more and more of the processes that are currently very manual and require multiple people and steps will become automated. Our customers will be right there with us as we help make those changes and see what that looks like. That can really change where health systems spend their time and do their work. I’ll give you an example. We announced this week what we’re doing with the Fleming Initiative, specifically around antimicrobial resistance. Every provider organization understands they have infections in their community, some of which are resistant to antimicrobial therapy, and they have whole departments of infection preventionists following surveillance and reporting. We’re now partnering with this initiative — antimicrobial resistance is extremely important. Right now, every eleven seconds someone is being diagnosed with a resistant infection, and every fifteen seconds someone dies from it. Being able to have a global platform connecting 150 countries, bringing together all of this information to minimize resistance — that’s going to change how they practice, how patients receive antimicrobial therapy, and those mortality numbers. When you automate practices that infection preventionists are currently doing perhaps on a spreadsheet, you free up that expertise to be deployed elsewhere, take away administrative burden, and help patients have better outcomes — not being treated with an antimicrobial therapy that isn’t working, or having to have a second course because the first didn’t cure the infection. Those are really significant leaps that are all about technology, automation, and using agentic tools to get to insights faster and change clinical care.

Ritu: That truly scales and multiplies impact to save lives. Thank you for sharing that, Dr. Shippy. I also wanted to ask about the Children’s Health Innovation Awards — I was reading about that and it sounded remarkable. Tell us more about those awards and what you’ve seen with the health systems you’ve worked with.

Angela: The CHIA — Children’s Health Innovation Awards — was an opportunity to say: here we have a population of children with rare diseases and genetic diseases, and children’s hospitals are standalone entities, often associated with academic institutions but smaller in scale than adult hospitals. Putting these awards in place was about helping them bring their information together — accelerating the genomic testing and genomic research they’re doing, accelerating work in rare diseases, and helping them come together on platforms where data from different parts of the children’s disease ecosystem — research institutions, children’s hospitals — could be aggregated. When you take all these smaller populations and aggregate them, you have more data that can get you to insights faster. We wanted to help them get to those insights faster, make an impact on providers taking care of children right now, and also shed light on the importance of rare disease research and genomics and what that can lead to in the future. We have a purpose-built service, HealthOmics, that helps get to insights faster from gene therapy and research.  I’ll share another example. Over five years our health equity initiative helped organizations look at underserved communities, access to care, and ensuring that patients — no matter where they lived, whether in a rural or underserved community — had access to technology. Over five years that was a $90 million commitment helping 600 organizations. Babyscripts, for instance, was able to diagnose pre-eclampsia faster — eclampsia is one of those true medical emergencies in a pregnant patient, so getting to diagnosis faster is huge. We were able to help a company in Uganda take care of at-risk babies for the six weeks after delivery, ensuring they were able to attend their postnatal visit — a 275% increase in postnatal visit completion, and a decrease in mortality as well. These are really significant impacts from funds we’re excited to give and support, either with technology embedded to improve the application or with technical support for applications being built on AWS.

Ritu: I hope listeners learn about the CHIA program and it reaches a bigger audience to help other health systems apply. Thank you for sharing that. As usual, the time has flown by, Dr. Shippy. Any closing thoughts you’d like to share? Where do you think technology is heading? We used to ask about the next three years, then the next year — now it’s almost the next three months because it’s moving so fast.

Angela: One of the most exciting things in our work with customers is watching them start small but think big, never losing sight of where they’re trying to get to. They’re trying to get to better patient experiences, better experiences for their workers, and ensuring that administrative burden is taken away so folks can work at the top of their license. Researchers really able to go fast, accelerate, and get to insights that move them from bench to bedside faster. Nurses and physicians at the bedside really able to interact with patients face to face, eye to eye — not looking down at a keyboard — because of ambient listening. Getting to the discovery of new drugs and new applications of devices faster. It’s really exciting to work side by side with customers as they do that. And the other thing becoming very clear — and our customers are starting to see it too — is that in healthcare, you want to be at the forefront of utilizing this technology. In being at the forefront, understanding that the next big discoveries are most likely going to happen with someone — whether an individual researcher, clinician, or an organization — that either has an agent by their side or agentic technology deployed across the organization. Whether it’s coming from the bottom up from frontline staff or top down from leadership, really embracing this technology and putting it to use to improve the experiences of everyone across the continuum of care is going to make a difference — both in the experience and in the overall health of all the populations being served. That’s really exciting as we look forward.

Ritu: Thank you so much for sharing all those insights with us, Dr. Shippy, and thank you for bringing this transformational change to so many people. It’s exciting to see you doing that. Thank you for being on our podcast today.

Angela: Thank you so much for having me, Ritu. It was great to be here.

—

Subscribe to our podcast series at www.thebigunlock.com and write us at info@thebigunlock.com    

Disclaimer: This Q&A has been derived from the podcast transcript and has been edited for readability and clarity.

Ritu: Hello, listeners. Welcome to Season Seven of the Big Unlock Podcast. My name is Ritu Oberoi, and I’m the managing partner at Damo Consulting and your host today along with Rohit, my co-host. We are really excited to welcome Greg O’Neill to our podcast. Greg is the Director of Patient and Family Health Education at ChristianaCare, one of the nation’s most innovative health systems, based in Wilmington. In this role, he leads the strategic plan for patient education and health literacy initiatives. I’ll ask Rohit for a brief introduction, and then it’s all yours, Greg.

Rohit: Hi everyone. I’m Rohit Mahajan, CEO and managing partner at Damo Consulting, and very excited to have this conversation with you, Greg. Over to you.

Greg: Thank you for that introduction. Just to clarify further — I’m an advanced practice nurse, so everything I do comes through that nursing lens. Nurses really try to connect with patients’ needs, as do other healthcare providers. I’m excited to be here. Thanks for having me, and let’s get into it.

Ritu: Why don’t we start with you telling us a little more about ChristianaCare, where your digital health program stands today, what the enterprise priorities are, and how patient education and health literacy fit into that broader strategy?

Greg: I’m very lucky to work at ChristianaCare. We’re a very busy health system — one of the busiest in our region, the largest healthcare system in Delaware, and we provide care across four states. A lot of people are depending on us to provide safe, effective care. We like to say that we serve our patients, families, and communities in their care. One of the ways we do that is by becoming more effective at communication and more efficient in care delivery. Patient education and health literacy are tools in our toolbox to meet people where they are and make sure they’re getting the care they need and can care for themselves as needed.  As for system priorities — we do like to be innovative and forward-thinking, and that’s one reason health literacy is so important to our organization. Technology is equally valued. While I don’t oversee all of the technology platforms, I can tell you there is serious interest in making things easier for patients, families, and caregivers alike. Everyone wants healthcare to be easier — for those providing care, they want everything to be seamless and connected; for those receiving care, they want to understand what’s in front of them, what their choices are, and what they need to do to have a better quality of life. That’s where we’re trying to meet them.

Ritu: Thank you, Greg. That’s a perfect lead-in to my next question. We’re seeing a lot of focus on digital access and the consumer experience because patient expectations are rising alongside access to tools like ChatGPT, where patients feel much more prepared and empowered when they come in for care. What are you seeing on the front lines, and how does that shift the mindset around care delivery?

Greg: Our entire world is undergoing this relationship with AI and information. The attention economy is very important — where are people spending their time and attention? One thing we can’t forget in all of this is that people spend time with information and sources they trust. So while we continue to evolve and make care more efficient using advanced systems like AI, we really need to make sure the human connection is still there. No care can happen with digital resources alone. You need care providers, a team, and a system as the backbone to support all of it. That human connection is really where, from a communication standpoint, it all starts and ends.  From my perspective, if a person leaves a healthcare encounter confused — not knowing what’s happening or what to do next — that will not help build trust between the two people in that interaction. The tools we use, like AI and patient education resources, need to be wrapped in the context of the human relationship to be trusted. Otherwise, when folks don’t fully understand what’s happening from their care provider, that’s when they seek other sources online. We really want to make sure they trust us first, so that the tools we’re using are what they turn to when they need additional information and want to follow up on their care.

Ritu: I have a follow-on question. What do you think patients and families are actually asking for when it comes to the digital health experience, and how has that shaped the way you approach medication communication and education?

Greg: There are two things I think about on that front. Everyone wants this to be easier — that is the main thing. And it has to meet people where they are. Those are the two pillars. If a person is interacting with an information source and the information is too complicated, too voluminous, or too complex, they’re not going to be able to access it. How we communicate is actually an access-to-care issue. If people don’t feel they’re getting information in the way they need it, they may disengage and disregard that information entirely. Going back to ease — we have all become accustomed to everything being fairly easy from a consumer standpoint. At the touch of a button you can have things delivered to your home or get the service you need, and I think that’s what people are looking for from healthcare: to evolve and meet the same lifestyle standard they experience in the rest of their lives. The goal is to make it that easy, but at a level that people can genuinely interact with.

Ritu: Absolutely right — they want instant access, but they also want to be able to trust the information they’re getting. Rohit, would you like to ask a question?

Rohit: Greg, the practice of medicine is so wide and specialized, with so many different subspecialties. How do you build trust across all of them? Is your group focused on one aspect, or how does a health system overall build trust with patients across every specialty?

Greg: That is an amazing question, because it’s extremely difficult. The care we provide is so complex, the needs so varied. At ChristianaCare, we take care of people from all stages of life — we’re with people the whole way. Real trust and connection happen at the local level. Whatever your interaction with a given specialty is, that’s where the actual connection to care takes place. But from a system strategy standpoint, my team oversees the strategic plan for how we conduct patient education across the organization. We do need broad-reaching standards and practices to ensure the baseline for everyone is at a good operating level. That’s why we promote health literacy best practices. We need to make sure the tools we’re using are wrapped in the care delivery practices of each local provider, so that local connection really sticks. And going back to ease of use — if we’re making system-level decisions that set a relatively simple baseline for everyone, that makes it easier for local providers to pick up those tools and use them in their individual patient interactions. It’s the combination of those two things — system-level standards and local connection — that really drives success in healthcare.

Rohit: That’s great. And Greg, tell us more about how you started your nursing career and what some of your plans for the future look like.

Greg: I started as a student nurse extern — a common way many nurses begin their careers, finding your way into a health system and getting to work right away. From those roots I moved through surgical ICU, where I was a nurse for a number of years, and then continued to advance within the system because I really wanted to impact as many people as possible. I try to bring systems thinking to my nursing practice. After the ICU, I went on to get my master’s and became an advanced practice nurse, because from that position you have access to roles where you can influence with broad strokes how we provide care — and that’s where my personality met my profession.

Rohit: What do you think the future holds?

Greg: I just want to keep providing value — that’s my main goal. If I can influence the conversation and the attention given to patient education and health literacy, I’ll keep doing that. There are many competing priorities in healthcare, and one of my roles is to be an advocate for patients through communication and health literacy best practices. I hope to keep elevating that conversation, because I think it may be underappreciated in the history of healthcare just how important communication is in helping people reach their health goals. Whenever there’s an opportunity for improvement, it often lives in the communication realm, and we need to make sure we’re connecting with people.

Rohit: I’m curious about your approach — communication is so diverse, with so many channels available. How do you measure success at the end of a quarter or a year? What are your success criteria?

Greg: In the realm of patient education, those connections are still being developed. When it comes to quality of care, we have a number of metrics that health systems everywhere use to monitor things — like readmission rates and hearing directly from patients and families about their experience of care. We try to align ourselves with those system-level quality goals. The belief is that by improving patient education and communication, we contribute to improving those standard metrics: safe and effective care, and hearing back from patients and families that it’s working. Specifically for patient education, we’re working to connect our tools more directly to the outcomes we’re seeing. That’s still evolving work.

Ritu: It’s time to talk about AI. We speak with a lot of C-suite leaders, but given your experience on the front lines of nursing and working directly with patients, what do you think about technologies like ambient documentation, AI scribing, or voice agents? How are they going to affect nursing, and where do you see the most transformational potential? Where are you seeing the best adoption and the most meaningful difference?

Greg: I don’t directly oversee systems like ambient listening, but I pay close attention to what’s happening in those spaces. To me, the potential is huge, and I think it’s really going to break down silos. It’s still a little early to see exactly how it will all come to fruition, but one of the most frustrating things for patients and families is when care feels disconnected. We do our best here to make sure those dots are connected for people. AI has so much potential to take what has evolved as disconnected systems and bring them together — to take the cognitive load off the patient, to consolidate all of this information into a clear understanding of what’s happening, what the plan is, and what they need to do next. If we can do that, we’re really transforming healthcare, and I do see a lot of attention moving in that direction.  Healthcare is not easy — I can say that directly from having been at the bedside as a nurse. Anything we can do to make the process easier for everyone involved, both caregivers and patients, is going to be a game changer. If we can make all the systems work together to support that, it will be a genuine lifestyle change for everyone involved.

Ritu: Like a personal companion that guides you through the whole journey and keeps telling you what’s next — that would be really remarkable. While looking at ChristianaCare, we read about your work with FDB and their Meducation solution to bring personalized multilingual medication calendars inside the EHR. For listeners or health systems that haven’t explored that kind of tool yet, what does it actually look like in practice?

Greg: When I started this journey over twelve years ago, really focusing on patient education and trying to bring health literacy concepts and practices to care delivery, I wanted to find a strategic partner. I knew this process was going to take time to evolve, technology would keep changing, and I needed a thought partner and strategic partner for the long haul. That’s what I found with FDB and Meducation. From the very first time I saw their solution, I felt they had health literacy at the forefront, and that really came through in the innovation of the tool. The whole thing was created to meet the need of people understanding their healthcare, specifically around medications.  For those who haven’t considered this yet, I would say: add a health literacy lens to any decision-making process you undertake. Think about plain language, ease of use, and how the tool will actually be used — and that will help you determine what’s best for your organization. I saw from the beginning a real prioritization of patient needs first. And my team, being nurses, we don’t want to make life harder for the nurses and providers going through their daily activities, so we focus heavily on integrating the tool into their natural workflow. As much as possible, we try to keep it within a step-by-step process that doesn’t pull them completely away from what they’re already doing. FDB has been a great partner in working back and forth to make the tool as easy as possible for end users, which in turn makes the end product for the patient feel meaningful rather than cumbersome — something worth doing for the patient’s benefit. The main commitment is to ensure that the process of using the tool remains easy enough for our caregivers.

Ritu: Healthcare leaders are increasingly being asked to do more with fewer resources. Beyond the financial ROI of digital transformation initiatives, how do you measure the impact of health literacy programs in terms of clinician satisfaction, patient outcomes, or system resilience? How do you justify the investment?

Greg: We all need to be making sure we’re adding value, and we strive to do that every day. While we’re still developing the data tools on the back end to connect those outcomes, what really resonates for me personally is the individual stories. Nothing makes a healthcare provider — nurse, physician, physician assistant, whoever’s delivering that care — feel more validated than when they genuinely feel the care they’re providing is understood and that the person is going to be able to meet their goals.  When we do training sessions, my team doesn’t just manage technology — we also manage best practices. When we work with nurses and physicians on how to use the tools and how to communicate clearly using plain language, you see the light bulbs go on. When a provider says, “It’s really important to me to communicate clearly, so I use teach-back in my practice — and when I hear it back from the patient, I’m validated,” that to me is the most meaningful metric. Everyone is trying to build trust and make the relationship the driver of care, and that’s the best way to do it.  So it’s both ends: we’re trying to build a system-level view on the back end, making those ROI data connections to quality and safety — but providers don’t necessarily come to work every day with data metrics in mind. They’re trying to provide excellent care, connect with people, and make sure care is valued and that people can meet their goals. The sweet spot is connecting the tools with the practices that hit home, so that when you leave at the end of the day, you know you really connected with people and made a difference in their journey.

Rohit: Greg, any closing thoughts for our listeners?

Greg: I appreciate the opportunity to keep sharing this message. I just hope people continue to prioritize communication and health literacy best practices. One thing we really want to avoid in healthcare is waste — wasting time or resources. The cost of care is on everyone’s mind, so we need to be more effective and more efficient. If we’re not communicating well, we might be spinning our wheels. All of our caregivers are very well-intentioned, doing the best they can. But if we’re not confirming that it’s working — if we’re not making sure people understand what’s happening before they leave our care — we might be missing the mission. I encourage everyone, whether you’re a systems thinker at the leadership level or a caregiver working directly with patients every day, to keep the health literacy lens on. If you do, you’re much less likely to miss that opportunity to truly connect with the people you serve.

Ritu: Thank you, Greg. It’s been a pleasure having you on the podcast.

Greg: It’s been a great time. Thank you so much.

—

Subscribe to our podcast series at www.thebigunlock.com and write us at info@thebigunlock.com    

Disclaimer: This Q&A has been derived from the podcast transcript and has been edited for readability and clarity.

Ritu: Hello, listeners. Welcome to Season Seven of the Big Unlock Podcast. My name is Ritu Oberoi, and I’m managing partner at Damo Consulting and your host today. We are very excited to welcome Umesh to our podcast. Umesh Rastogi serves as General Manager in Microsoft’s Healthcare and Life Sciences Dragon and Platform Group. In this role, he leads product and engineering teams developing innovative AI solutions for nurses, as well as broader data platform capabilities for the healthcare industry. Umesh is deeply passionate about building technology-driven innovation to help the healthcare industry achieve its quadruple aim, and he has held pivotal roles in engineering, technology, architecture, and product management. Welcome, Umesh, to our podcast. Really happy to have you here.

Umesh: Glad to be here, Ritu.

Ritu: Would you like to add anything to that introduction?

Umesh: My journey has been more than two decades in enterprise software. I’ve held a variety of different roles and worked across both the applications domain and platform. My foray into healthcare was about ten years ago, actually when I was helping my daughter with her biology coursework — that’s when I thought about how I could use technology to have an impact on healthcare. Over the years I’ve had the opportunity to build hospital emergency response solutions, and when COVID hit, various other healthcare capabilities along the way. And as you mentioned, I’m now leading our product and engineering teams building the Dragon Copilot solution for nurses, and I’m very passionate about the work we’re doing there.

Ritu: Thank you for sharing that story. We always love origin stories, and we find out the most interesting things about people. Let’s get started. Healthcare leaders often talk about AI as if the primary challenge is just selecting the right model or the right LLM. But much of your recent work has focused on building unified multimodal healthcare data foundations. Are we at risk of overestimating the AI problem and underestimating the data problem? What does AI-ready really mean in a healthcare organization?

Umesh: That’s an excellent question. If you take a step back and ask what outcomes these organizations are typically trying to achieve, it fits within the quadruple aim: improve healthcare outcomes, improve patient experiences, reduce cost, and improve the well-being of clinicians — physicians, nurses, and so on. Whatever technology we pick has to address one or more of those aims. Many organizations end up making it purely an LLM or core model problem, but that’s just a piece of the puzzle. The whole technology stack actually starts from making sure you have access to the right data — the data on which you’re applying AI models and agents and enabling workflows for physicians, nurses, radiologists, and others.

Ritu: Tell us more about how Microsoft is partnering with health systems to responsibly scale AI-enabled clinical solutions specifically for nursing workflows.

Umesh: If you look at what’s happening in the nursing workforce, recent data points to a shortage of about 80,000 nurses as of last year, and that gap is expected to increase to about half a million by 2030. What’s driving that is obviously the increasing demand from an aging population, but more importantly, high turnover within the nursing workforce — on average about seventeen percent across health systems in the US. The primary driver of that turnover is that forty percent of nurses report high stress and burnout. And the biggest contributor to that stress and burnout is that they spend forty percent of a typical twelve-hour shift in the system rather than taking care of patients.  The administrative work in EHR systems includes flow sheet documentation — rows and rows of data to capture as part of patient assessments, including vitals, head-to-toe assessments, patient safety documentation, and so on. That’s the biggest chunk of that forty percent. On top of that, nurses spend time looking for information — either in the EHR about the patient chart, or looking up drug adverse effects from external credible sources, or searching their own internal protocols and guidelines. And they handle various communication and coordination tasks, many of which could be delegated. All of this forty percent drives high cognitive load, turnover, overtime, and poor documentation quality. That’s the challenge nurses are facing today.  What Dragon Copilot is doing to address this is somewhat unique. If you look at the differences between physician and nursing workflows, nursing is typically much more fast-paced and mobile. In an ambulatory setting, a patient comes to visit the doctor — it’s a different environment. In nursing, it’s a very fast-paced environment: nurses are moving around, taking care of patients, attending to a given patient multiple times during a shift. The workflows had to be rethought. Over the last two years we’ve been co-innovating with nine different organizations — including Mercy Health, Advocate Health, and Stanford — learning, validating, and iterating on the solution to make sure we get it right. We knew from day one that for nurses to actually use and like the solution, it had to be as low-friction as possible and naturally fit the workflows they were already accustomed to.  We invested heavily in usability, in making sure AI accuracy was right, and in thinking through what organizations would need to scale the solution to thousands of nurses — because this is an entire workforce transformation. The result is what I would call the most comprehensive AI clinical assistant for nurses, which not only streamlines documentation but also surfaces information and automates tasks nurses currently handle manually. The nurse carries the app on a mobile phone, and as they see the patient at the bedside, they select the patient record and have a natural conversation with the patient. Or if they’re moving between patients and have an ad hoc observation — perhaps something they didn’t want to say in front of the patient — they record it and leave the rest to AI, which does the necessary extraction and automatically populates the flow sheet. Nurses can quickly review, edit if needed, and file. In addition, through Dragon Copilot’s chat functionality, nurses can ask about past patient interactions or access pending care activities, because Dragon Copilot provides a kind of cognitive memory of everything that has happened with the patient across the shift. The net result is a meaningful reduction in cognitive load.

Ritu: Thank you, Umesh — really detailed and well-explained. One of the great points you made is that the solution has to fit naturally within existing workflows and eliminate friction, because that’s what leads to so many technologies failing at the POC stage and never scaling. That leads into the next question: when nurses think about AI, there’s often resistance and skepticism. What have you actually seen resonate once nurses experience this AI firsthand? Are you seeing better adoption rates and a more open mindset toward the technology?

Umesh: There is a lot of hype and misinformation in the industry around what AI is and isn’t and how it will change things. From our standpoint, nurses who have actually experienced this technology firsthand tend to share feedback in two main categories. The first is around their own cognitive load. Nurses are saying the system is now doing a lot of things for them. Without this technology, when a nurse is doing documentation away from a workstation, they’re often scribbling observations on their hands to key in later, or trying to remember what they told a patient and whether they followed up. That adds to cognitive burden. Now, because the system ambiently listens to the conversation with the patient and records everything, nurses don’t have to carry that mental weight. The system has already extracted the information. And importantly, those conversations often contain more than just flow sheet data — for example, a promise to follow up on something. The system surfaces that as a pending care activity, and provides a running summary across all interactions with that patient during the shift. These capabilities significantly reduce cognitive load.  The second big category is presence. Nurses are saying they now feel present with the patient. Previously, nurses might have their back to the patient while keying things into a workstation, or be mentally focused on documentation while asking questions — not truly present. With this technology, nurses report being fully there with the patient, with a much freer flow of communication and needs, without worrying about explicit documentation. They leave that to the system, which does the extraction automatically, ready for review and filing. Those are the two main categories of feedback we hear from nurses who are using the solution.

Ritu: This is what we’ve heard from other C-suite leaders as well — an indirect benefit is that patients also feel more heard, because nurses are articulating more. Since the ambient system is capturing everything, nurses explain things more fully, and that benefits patients too. As generative AI becomes more embedded in healthcare workflows, there’s increasing discussion about responsible AI, governance, and trust. How should health systems balance the urgency to innovate with the need to ensure that AI-generated insights remain transparent, explainable, and clinically trustworthy — especially as the volume of information makes a human-in-the-loop approach harder to sustain?

Umesh: I would identify three aspects. First, before we ship any product or AI capability, responsible AI is front and center in Microsoft’s product development process. Nothing gets released without going through all the necessary checks — bias evaluation, review by nurses on our own teams who can assess whether an answer makes sense, and all the compliance requirements. That evaluation is done before anything is released.  Second, from a workflow standpoint, Dragon Copilot — whether for physicians, nurses, radiologists, or other clinicians — functions as a copilot. The user gets to review the AI output, edit if needed, and then file. Yes, AI can sometimes be wrong. And in these specific workflows, incorrect documentation could create patient safety risk, so clinicians are always expected to review the AI output before filing.  The third aspect: many organizations are saying they can’t treat AI as a black box, and they want transparency into how it’s behaving. We’ve built capabilities within our product to address this — we believe in establishing trust through transparency. Customers can configure their system to export raw data into Microsoft Fabric or their own data lake, giving them access to recordings, transcripts, and full visibility into how the AI actually performed. Organizations use this for audit purposes, for research, and for training. Based on ongoing customer feedback, we continue to develop additional capabilities to provide even more transparency — including visibility into usage patterns and what kinds of outputs end users are receiving.

Ritu: You mentioned co-innovating with nine large health systems and learning from them. Given all that experience, what are the key lessons for implementing at enterprise scale? What would you caution health system leaders about, and what would you want them to take away?

Umesh: There were two distinct categories of learning as we went through this experience. The first was iterating on the technology and the core product. Through working with nurses, nurse informatics leaders, and nurse leaders, we enhanced AI accuracy, improved the usability of the solution, and added functionality to continuously reduce friction in the workflow. We also built the organizational tools needed for a mass rollout. Pilots were manageable — a few hundred users. But a full-scale rollout means thousands or tens of thousands of nurses. You need visibility into actual usage across your workforce, across nursing units. So we invested in analytics tools giving nurse informatics leaders access to usage data and patterns. We also built AI tuning and customization tools, because each organization — and even each unit within an organization — has its own flow sheet schemas and charting practices. The AI solution has to align with those specific schemas and practices. We gave customers and their nursing informatics teams the ability to tune the system accordingly.  Almost equally important was the second learning: how these solutions actually get rolled out within organizations. At the end of the day, this is a workforce transformation and a change management challenge. It’s a different way of doing things — essentially building a new habit for nurses. Not every nurse is comfortable speaking aloud or charting aloud. We found demographic differences: more tenured nurses tend to be more comfortable with it than Gen Z nurses, which was somewhat counterintuitive. These non-product learnings about how the technology needs to be rolled out are critical to achieving maximum adoption. Change management is key — ensuring nursing leader sponsorship, nurse manager sponsorship, and dedicated champions for the technology in every nursing unit. We encapsulated these as change management best practices so that organizations can use the same technology in the best possible way for their specific context and actually achieve adoption.

Ritu: Digital literacy is so important — it helps remove the skepticism we talked about earlier. Once people understand the technology, they become much more open to adoption. We’re almost at the end of the podcast. Looking ahead one to three years — we used to say five years, but that horizon is now too far — do you believe the biggest impact of AI in healthcare will come from automating administrative work, enabling personalized patient engagement, or do you think an entirely new kind of care model will be unlocked?

Umesh: Restricting to the next three years and focusing on how nursing workflows and care delivery will change — I think we’ll see smart hospital rooms and wearable devices become far more integrated into workflows, creating a much more frictionless experience for both nurses and patients. Today we talk about audio; I think audio plus video will become integrated as well. I believe we’ll build a kind of AI-native UI that can pull together and surface information from far more data sources — delivering not just information but genuine insights. We already have early success there: bringing together data from the EHR with external credible sources and internal guidelines documents. Lastly, I think we’ll see a lot more agentic workflows rolled out. Today we talk primarily about streamlining documentation, but there is so much that nurses handle in interactions with supply departments, transportation, facilities, and so on — all of which takes time away from direct patient care. We’ll see more of those non-patient-facing activities delegated to agentic execution. I think those themes will come together very neatly over the next couple of years.

Ritu: Thank you, Umesh. Thank you for being on our podcast today and sharing all your learnings with us. I hope our listeners take away a lot of valuable insights. On behalf of the Big Unlock Podcast, thank you for joining us.

Umesh: Thank you, Ritu. It was great to be here.

—

Subscribe to our podcast series at www.thebigunlock.com and write us at info@thebigunlock.com    

Disclaimer: This Q&A has been derived from the podcast transcript and has been edited for readability and clarity.

Ritu: Hello, everyone. A very warm welcome to all our listeners to The Big Unlock Podcast Season Seven. My name is Ritu Oberoi, and I’m the managing partner at Damo Consulting and your co-host today along with Rohit. We are really excited and honored to have Dr. Haider Warraich on our podcast today. Dr. Warraich is a cardiologist, researcher, and one of medicine’s most compelling contemporary writers, known for bringing humanity, history, and philosophy into conversations about health and illness. He joined ARPA-H as a program manager in December 2024. ARPA-H is the Advanced Research Projects Agency for Health, which supports the development of high-impact solutions to society’s most challenging health problems. He was previously a faculty member at Harvard Medical School and Brigham and Women’s Hospital and director of the Heart Failure Program. He has written three books and over 175 peer-reviewed papers. Really happy to have him here with us. I’ll ask Rohit for a short introduction, and then it’s all yours, Dr. Warraich.

Rohit: Hi everyone. I’m Rohit Mahajan, CEO of Damo Consulting and co-host of The Big Unlock Podcast. Very happy to be here. Over to you, Haider.

Haider: It’s my pleasure to be here, Ritu and Rohit. We first met at HIMSS and I know we’ve been meaning to speak since then. Lots of exciting things are happening in the health technology space, so I’m excited to dig in.

Ritu: Great. Let’s get started. ARPA-H was created to tackle ambitious, seemingly intractable challenges with a very different risk tolerance than traditional research funding. Based on your experience there, which healthcare problems do you believe are truly moonshot-worthy? And are we being bold enough in the questions we’re asking, or are we still being cautious and not fully exploring the potential of AI?

Haider: To create a thriving funding infrastructure, you need a variety of payers and funders who can tackle different bottlenecks in the market. You need an agency like NIH with broad scale where progress is realized over decades, if not longer. At the same time, private funders — especially in the AI space — will play a huge role in supporting breakthroughs and promising teams. ARPA-H fills a gap that existed in federal research funding: there was no health-focused agency that could take big swings, turn things around quickly, operate on a metrics-and-milestones basis, and have the risk tolerance to take on areas where incremental progress will never be enough — areas of genuine national interest that also create common good for the rest of the world. Our agency director, Dr. Alicia Jackson, has outlined key priorities for the agency. These include winning the biotech race, longevity and reversing aging, defeating devastating diseases, and taking a platform approach to AI. Within AI, we have many different programs approaching this from different angles. The program I’m leading is called the Advocate Program, and its vision is to bring AI to the bedside — bring generative and agentic AI that can essentially do everything a clinician can over the phone — and take that through the regulatory pathway so it can open up this entire field for everyone. But to do so not just quickly, also done well, given how critical it is to get this right. The response to not just this program but to many of ARPA-H’s programs is a testament to how important it always was to have us as part of this ecosystem.

Ritu: I want to dig a little deeper into the speed at which AI is moving. When you take on a project like Advocate with a six-month horizon, but everything can change in a week — how do you keep up and think that far ahead?

Haider: It’s a great question, and it’s actually what makes developing a government-funded AI program very difficult. Even the most ambitious projections in this field can become obsolete very quickly given how fast things are moving. Here’s how I’ve approached it. First, there’s a reason Advocate is focused on the highest-risk AI in the most regulated space. There is already significant integration of AI across healthcare, but mostly in low-risk settings — ambient scribe technology has taken off, and clinical decision support will transform how clinicians interface with the healthcare system in the next few years. But Advocate was focused on developing a patient-facing technology, which puts you in a high-risk but also high-reward category. It targets really high-capability functions: making a diagnosis autonomously, performing triage, closing the loop by providing and prescribing treatment directly to the patient, and going through a regulatory pathway where there’s essentially no precedent for technology like this. We also focused on a relatively high-risk population — patients with heart failure — because ARPA-H believes that if you create a demonstration this ambitious, it opens up the field for everyone downstream. The reverse doesn’t happen. I also tell people: if you’re developing an AI program and it doesn’t sound insane to you, you’re going to get scooped. If what you’re doing sounds feasible today, you’re already a few steps behind. A lot of the work we’re developing for the future would sound off the wall if I described it today — but that is the kind of ambition we need to bring to this technology, because so far it has delivered on, if not exceeded, expectations.

Ritu: How do you reconcile that supersonic speed of development with regulatory pathways that are still slow? If it takes months to approve something, you risk falling behind again.

Haider: That is one of the reasons why an agency like ARPA-H should — and does — work very closely with the regulator. The FDA has never had a partner in government like ARPA-H before: one that can make major investments and ensure that the problems we’re solving actually help the regulators, rather than just impeding their work or getting in the way. That partnership is really crucial. I do think that many regulatory unlocks are fundamentally bureaucratic issues — a sense that you need X number of patients, a specific margin, a specific enrollment type. But in the AI space, a lot of the regulatory blockers are technical. We simply don’t yet have the technology infrastructure that this space requires. One of the most important parts of the Advocate Program is not just supporting the development of technologies that will be classified as medical devices by the FDA, but also funding evaluation technology for post-market monitoring. We call this a supervisory agent, or the supervisory layer — and it is being developed at the same pace and with the same level of resources as the primary technology, if not more. Rather than building in sequence, we are building in parallel, because we understand that traditional timelines and pathways are not going to work for this type of technology. If you want to be relevant as a body that can administer regulations in this space, you need technical support. We have to funnel and focus the best minds and best teams in the world on the problems that are going to unlock this field — and working with the regulator to keep their perspective centered ensures we create solutions aligned with their strategy. We plan to keep doing that.

Ritu: One of the most interesting things I picked up is that the Advocate Program is patient-facing. When you’re talking about technology that has to sound almost crazy to be relevant, how do you bring it to people who may not be ready for it?

Haider: There are two things. First, when you start with an ambitious vision, you also need to scope it. Instead of saying the teams Advocate supports are going to build AI cardiologists, what we’re focused on is: can you develop a technology using generative and agentic AI that improves outcomes in patients with heart failure? That immediately scopes the problem — you define the population, the endpoints, what goes on the label, how you design and deliver it. Scoping is very important. Second, communication matters enormously in the AI space — more than in almost any other field — because people genuinely care about this. This is not a technology working quietly in the background. Everyone sees AI as something that could transform society, healthcare, and every industry. That creates excitement, but it also creates anxiety. There is a real risk that poorly developed tools could add more friction to a system that doesn’t have a lot of give. Technology that isn’t well designed, or doesn’t have the patient’s or clinician’s best interest at its core, could be genuinely catastrophic. So how we talk about this technology and how that is reflected in the design matters enormously for getting the market ready. The technology is ready — the market needs more time. But the market also needs to see things in familiar forms. Clinicians take risks with new technology all the time — new implantable devices, new drugs, new chemotherapies, new genetic therapies — but there is always a sequence to what they see before taking that risk: clinical trials, FDA authorization. There is a precedent for what a community needs to see before accepting that the benefits of a new technology exceed the risks. That is what the first generation of clinical AI technologies needs to provide. For patient-facing technologies, the bar should be, and I think it is fair to argue it should be, the same as for a new high-risk medical device or a new drug. Once that foundational evidence exists, it will cause a seismic shift in how these technologies are accepted, implemented, and deployed. That is one of the core theses of the Advocate Program. Even though the technology itself is frontier — agentic AI for clinical care, supervisory agents for real-time post-market monitoring — the evaluation is super traditional: single-site deployments, human factors research, IDE studies, multicenter trials, and outcomes people understand: heart failure hospitalizations, guideline-directed medical therapy, mortality. We chose to innovate in the technology but kept the evaluation as familiar as possible, because that is the evidence people are accustomed to seeing when adopting new technologies.

Ritu: Great answer. Now I’ll give Rohit a turn.

Rohit: Thank you, Ritu. Haider, I especially liked the point that the FDA has never had a partner in government like ARPA-H — that’s fantastic, because regulatory hurdles need to be surmounted quickly. Tell us a bit more about the Advocate Program: the time horizons, the funding size, the number of projects. And separately, tell us about your own story — what attracted you to healthcare and to ARPA-H, and what you’re looking toward in the future.

Haider: On the timeline: the Advocate Program was launched in January. We solicited solution summaries from teams across each of the technical areas — focused on the development, evaluation, and implementation of clinical agentic AI, patient-facing, for patients with heart failure. That is standard process for every ARPA-H program. We then encouraged teams to submit full proposals, from which we make decisions about who to recommend for funding. Our hope is to announce the selected teams in the coming months, sooner rather than later. On funding: these are contracts, not grants, and a typical ARPA-H program is usually between fifty and a hundred and fifty million dollars. That said, the value the program was always created to provide goes beyond what you can count in dollars. If you’re going to make a dent in the AI space, money alone is not enough — there is plenty of capital available. The real question is how we create value and something that genuinely matters. We are leaning into our strengths as a government agency, which means focusing on the regulatory pathway, on implementation, and on helping health systems get excited about technology where there are still a lot of unknowns. Healthcare systems wanted a partner they could trust as a convener of some of the best technology companies working in this space. And then there is access to data — data that will help determine how well these technologies are developed and how well they perform after deployment. Our goal for the teams is a technology ready to go to FDA for market authorization within two years of the program starting — one year felt too short, three years felt too long. After that, we plan a multicenter trial of the highest-performing technologies in an integrated health system, focused on clinical outcomes but specifically on how to implement this technology in clinical practice. We hope to share more details in the coming months.  As for me: I grew up in Pakistan, not far from where Ritu is from in Delhi, just across the border, and went to medical school there. I have always approached writing as a storyteller first, and that has informed my research and clinical career. I moved to Boston for residency after finishing medical school, completed a cardiology fellowship, and then had the opportunity to join the FDA as a senior advisor on chronic disease to the FDA Commissioner. That experience brought everything into sync — giving me the regulatory perspective, helping me understand the critical bottlenecks, and helping me find a common language that works for technologists, private investors, academics, health systems, and regulators. I have been at ARPA-H for almost a year and a half, and it has been really exciting.

Rohit: Thank you for sharing that. You mentioned the Advocate Program application cycle has closed — are you following a January-through-December cycle, and will the next cycle open in January for teams that want to know?

Haider: The Advocate Program opening has closed and we’ll be able to share more once all the teams are on contract — that is essentially day zero of the program. We’re hoping that comes soon. Overall, things are going well. The way ARPA-H works, program managers are expected to develop new programs roughly every year, ideally building on and extending the thesis of the previous one. I foresee — and I’m hopeful — that you’ll see our team coming out with more programs that build on Advocate and extend its impact across the field. This space is too big to rely on a single program to unlock it. Teams are also always welcome to send their ideas directly to ARPA-H and connect with program managers. That process is competitive and depends on finding the right idea, the right risk space, and the right team, but it is an opportunity that remains open. These investigator-initiated projects, alongside agency and program manager-initiated programs, are the major ways to engage with ARPA-H, and there is always new work and new opportunities coming up.

Ritu: I think we can squeeze in one last question. Seeing this field so up close and being in touch with teams telling you about the future of healthcare innovation — which emerging technology do you think is receiving more attention than it deserves? Where is everyone focused when there are other things we should be looking at?

Haider: My paradoxical answer is probably AI. What I’ll say is that AI is receiving the appropriate level of attention, but there are still real gaps we need to focus on. How high-risk AI technologies are evaluated — real-time evaluation — is going to be incredibly important and needs more attention. AI alignment is another: how do we align AI outputs with the preferences, values, and needs of humans, not just as groups but as individuals? That is especially important in healthcare. I also think AI developed specifically for the populations with the greatest need — older people, people with multiple comorbidities, cognitive dysfunction, sensory issues — needs more focus. Developing and validating tools for those populations, where we think the benefit can be greatest, is underexplored. And finally, the space I’m most personally interested in: I think we have finally reached a point where we have the models, the data, and the compute to develop high-fidelity simulations for healthcare. In every other field — manufacturing, supply chain, logistics, transportation, defense, climate — simulation is a huge part of how innovation happens. It remains essentially nonexistent in healthcare, largely because healthcare is hard, our data is messy, humans act in complex ways, patients have intricate physiologies and pathologies, and the overall complexity has made realistic simulation essentially impossible until now. That is a space I believe has real potential to transform how we develop and value technologies, and could meaningfully shorten both development and implementation cycles.

Rohit: Haider, would you like to say anything to close out — perhaps about your most recent book or anything you’re working on?

Haider: Before joining government I had written three books over five years, and in some ways joining ARPA-H was my retirement from writing — I wanted to take a break and be fully present in this work. My last book was focused on chronic pain, partly because from a scientific perspective it is a fascinating phenomenon, partly because I practiced at the height of the opioid epidemic in the United States, and partly because I have lived with chronic pain myself — I’m all good now, but for many years it was part of my life. I combined all of those threads into a book called Song for Scars, published around 2021. But it has been a while since then. What I will say is that a lot of the work we do at ARPA-H is unexpectedly similar to writing a book — having a plan, an outline, a vision, and then putting the pieces together, talking to people, doing research, iterating, editing. It is a very creative, collaborative process. So even though I thought I had retired from writing, much of what I learned then has proven deeply applicable to developing programs and working within this amazing ecosystem.

Ritu: Thank you, Dr. Warraich. It’s been a pleasure having you on our podcast. Thank you so much for joining us.

Haider: Thank you. I appreciate it.

—

Subscribe to our podcast series at www.thebigunlock.com and write us at info@thebigunlock.com    

Disclaimer: This Q&A has been derived from the podcast transcript and has been edited for readability and clarity.