Investing in AI-Native Healthcare with Venrock's Bob Kocher and Unity Stoakes

0:04Bob Kocher: Why did they quit their day job and start this company? I mean, there's always a really fascinating story because every founder I met was doing something different before they started their business. And it is a scary, giant leap of faith to go start a company and it's like having a child. You have to bring it to life and bring people to it. It's like the hardest job that exists, I think, is starting a company because we have to make something out of nothing. So understanding why. Like, why was this the problem that they quit their proper job and spent their life on is like really interesting to me.

0:38Unity Stoakes: My name is Unity Stokes, co-founder of Startup Health. I'm really happy to welcome Dr. Bob Kocher back to Startup Health. We go back a long way, I think over 15 years. In fact, Bob was there when we launched Startup Health at Health Data Palooza. I think in 2011. Right about then, yeah. And you taught us something very early on that has stuck with us ever since. In fact, we've designed it right into everything we always coach to founders ever since, which is in healthcare, the person who pays, the person who buys, and the person who actually benefits are often three completely different people.

1:24Unity Stoakes: If you're building a healthcare company, understanding that dynamic can be really the difference between building something people love and building an actual business. Bob has one of those rare vantage points on healthcare. He helped shape national healthcare policy during the Obama administration, helping architect the ACA. In fact, he spent years at McKinsey, trained and still practices today as a physician. And really today is one of the world's great healthcare investors as a partner at Venrock, where he's backed and helped build some of the most important companies in healthcare, including Devoted Health and Virta Health.

2:09Unity Stoakes: Companies we're very proud to have invested in at Startup Health. So today is really our opportunity to learn from Bob. And we hope, As always, you post your questions in the chat. After Bob and I have a conversation, we'll open it up for questions. Ask your questions early, and I will do my best to call on you and welcome you to the conversation. Bob, great to have you back with the Startup Health community. Welcome.

2:38Bob Kocher: Unity, it's great to see you. I so appreciate the work that you and Startup Health do. You've created a tremendous movement in the ecosystem and so many great companies. It's always been to talk with you and I'm glad to be back to talk with you again.

2:50Unity Stoakes: Thank you so much for that. So to kick things off, I thought we'd start macro to better understand what world are founders actually building into? How would you describe the landscape from your vast vantage point today?

3:08Bob Kocher: Well, I think we're building into a great opportunity, which is nearly every aspect of healthcare in America can be made better. access is a giant problem. The quality can be greatly improved and affordability is really a critical issue for us to tackle as a nation. With the magic of AI, I think we can make things faster and better than we could before. And in many ways, I think it's going to become tougher because with HR1 going into effect next year with many people losing Medicaid, it's going to create a lot of demand for lower cost ways to provide care and a bunch of disruption. So I think we're at a moment where, like maybe we've been for the last 15 years, it's never been a better time to start a healthcare company than today.

3:48Bob Kocher: And that remains true. I'm also pleased that the funding environment, I think, is positive. So if you have an idea that TAMs are so big and the demand is so great, and this is kind of a counter-cyclic area, then I think it's a good time to start a business.

4:01Unity Stoakes: We couldn't agree more as these windows of opportunity sometimes open up, as they did at the dawn of the internet and world wide web and mobile and other moments, ACA being one. One thing I think is interesting from things I've heard you speak about or write about recently is you have a somewhat contrarian argument right now about AI that while it may make healthcare better in so many ways and more accessible, but not necessarily cheaper. Tell us more and what you think that means for innovators today who are building?

4:38Bob Kocher: Well, AI is going to make healthcare better for sure. As a doctor, I can tell you AI is better at evidence-based medicine than I am, and I think probably any doctor because it knows all the evidence and it updates all the time. So it's incredible. You can engage patients in dialogue. Patient-reported outcomes are easy to capture. I mean, it's amazing. But AI will work in the system and system incentives that exist for the people in healthcare. And since most healthcare today is fee-for-service, most people using AI are going to be using it in a fee-for-service context. And AI creates productivity.

5:10Bob Kocher: AI can allow you to do more virtual care. With scribing, see more patients a day because you don't take notes. And so it will create more visits and visits are expensive. And so since most of healthcare is fee-for-service and AI is going to create productivity, it's going to lead to more activity. It's going to be more expensive. Now, it might be good. It might improve access. It might improve quality, but I don't think it by itself will lower cost of care. We have no mechanism to actually lower prices in our system if the care gets more efficient. And so I think it's going to not have a deflationary effect.

5:39Bob Kocher: In value-based care, it's going to help people in value-based care do that care more cost-effectively. And again, they'll make more money in value-based care, but it won't lower the prices or lower what people pay. Because outside of Lasix, I know of no examples of how the price is falling. Lasix and generic medicine participants.

5:57Unity Stoakes: It's great that there's more access, but we still got to solve the cost.

6:02Bob Kocher: I was talking to a large employer who was talking to me saying, Bob, I wish my employees would all get permacore doctors and go to the doctor. And I said, I would like them all to have permacore doctors. I think it's good for people to have a way to access healthcare. And then they said, half my employees don't go to the doctor in a given year. And if they all went to the doctor, we would spend a couple hundred dollars more per employee and that would be more expensive. And so then you kind of have a debate about, should they? Like, is that good or bad? And that's a good debate to have. And is primary care lower cost or like, and the answer is maybe, but it doesn't make it cheaper this year.

6:39Bob Kocher: It may make it cheaper in a couple of years because primary care will identify things and treat them. But in the year that you begin care, it may be as cheaper. That's why preventative care also was so tough, is that I can help all of us, I think, you know, sleep better, eat more healthy, be more active, stop habits that are dangerous. And that'll be good for us. But it costs something to do that. And the payback may not be for a while. And our health system thinks in one year increments. And so it makes it really hard.

7:06Unity Stoakes: Yeah, I want to dig into that in a minute, because I want to tackle some creative business models around prevention, something near and dear to both our hearts for so long. But as you really look across the landscape today of healthcare, are there two or three places where innovators or startups can still create enormous value? Or I guess, where would you advise them to focus first to create that enormous value over, say, the next five years? Is it?

7:37Bob Kocher: I go back and say, I think almost everything can be made better. And the Rams are big everywhere, so I think there's a lot of spots. Maybe do the inverse. Where would I not do it? I think what we're really good at in America is intensive care in hospitals. So I think ICU care is hard to do better than what we're doing today at the hospital. I think, frankly, surgery centers and operative procedures are done awfully well and quite efficiently, so I probably wouldn't start there. I probably wouldn't do generic telemedicine because I think there's so many players in that space. So that would leave you with everything else.

8:10Bob Kocher: Nobody's figured out value-based specialty care. And so that remained an area I think a lot about how one might do that. I think geriatric behavioral health has done very poorly. And so there's probably ways to do that a lot better. I think medication adjustments are often not done very effectively. A lot of people are taking medicines for their blood pressure, for their cholesterol, as examples, and they're not controlled. And so there's probably a way to improve the titration and usage of those medicines. And then deprescribing is one thing that devoted that we spent a lot of time figuring out how to do properly because people are on medicines that no longer are needed.

8:43Unity Stoakes: Yeah. And I know you've also talked about moving healthcare from episodic visits toward more of a continuous care.

8:52Bob Kocher: I mean, it'd be better not to have me see you once, twice, or three times a year, but to actually be in dialogue with you. So when you have a question, you can ask it. Now with AI, that's much more practical, but obviously things like blood pressure control, it's silly to come back to an office every eight weeks and have somebody measure your blood pressure once and be like, okay, we're going to change it. You should just check it multiple times a day and we can change the medicine tomorrow. We should just get these things done a lot more. That's why CGMs for diabetes is so much better than, again, a finger pick a couple times a day and going to the doctor in three months.

9:21Bob Kocher: We should be able to do a much more rapid adjustment of medications.

9:25Unity Stoakes: So do you see the winners there being not primarily software opportunities, but more, they have to own some portion of the care delivery as well to really

9:37Bob Kocher: I prefer having more control over the care delivery. There are going to be autonomous software things also that could be interesting. Updoc, as an example, is a company that has an FDA approval due to titration of medicines for diabetes and soon probably other things. And so there could be software only, but I think it's easier still to have a clinician who can just change in the order for a medicine so you can close the loop.

10:01Unity Stoakes: Yeah. And, you know, obviously, And you started to go there on this. There's the big debate for many years about AI replacing doctors or pieces of the clinician versus augmenting the clinician. I think we're not replacing most doctors.

10:18Bob Kocher: I think we're augmenting. AI will allow you to be in that constant dialogue. It can allow you to not have to. Like in the past for remote patient monitoring or chronic care management, we'd spend a lot of time trying to stratify patients to make sure we did it just for the right ones where we'd have an impact. And so we can only do it for high risk people, but then we miss people who could benefit because it was cumbersome and costly and hard to do. With AI, I can offer that to anybody potentially. And we can find stuff and educate people and engage them in their care. And that's probably really good for them and good for care.

10:48Bob Kocher: And so I think it's nice to have like software can make a lot of things more possible, but I think it augments the clinician because ultimately care works better, I believe, when you have a patient goal that you and the patient are working to achieve. and when you have software it may treat every disease as if it's the most important thing in your life and that may be as impractical if you have more than one disease or a life

11:10Unity Stoakes: yeah totally and obviously so you're still seeing patients periodically and so what's something in your actual clinical workflow where you think i can't believe someone hasn't solved this yet

11:26Bob Kocher: god i mean so many things well first i think describing is the greatest thing since i spread because i don't have to type notes or dictate into dragon anymore so that's magic but one thing that's super annoying i had to do it today so it's top of mind is credentialing every two years or three years depending on the insurance company in the hospital you actually have to resubmit all this paperwork to get recredentialed and i cannot believe like i'm still sending pdf pictures of diplomas and things. It's just, it's, it's shocking. And like, there seems like there's so many ways one can actually, that they should know all this information because it's easy to know if I've been sued and if my licenses are active and if my CME is done.

12:04Bob Kocher: And so that's one where it just drives me nuts. I was just answering a question about sending a copy of a certificate that I'm like, this from 22 years ago. Are you serious? It was, you had it three years ago and it was fine. So why are we asking for it again? That's super annoying. I think, you know, referrals are still more complicated than they should be. Getting a question formulated and the information shared with the clinician and the patient scheduled and the information back, that's cumbersome. There's all these tools that aren't connected to actual practice around how to do this more optimally, but it's very hard in practice to use that.

12:39Bob Kocher: I think decision support, every note should automatically be bounced off of an evidence-based resource for the open evidence to check it for accuracy. It doesn't do that automatically. I think there's so many ways to make it better for sure. Yeah.

12:55Unity Stoakes: And so let's talk a bit about what you think is actually investable. So keep that hat on. You probably see hundreds of companies, for example, claiming to be AI native or claiming to reduce costs or whatever the case may be. What really separates, let's start with the AI, I guess, since it's where all the money's going, what separates an AI healthcare company you might seriously consider looking at or investing in from just something that's a wrapper or a feature? What has to be proprietary? What makes something investable, I guess, is the real question.

13:32Bob Kocher: I think if it's an AI thing, it comes down to distribution. So unfortunately for AI creators, anything that you create in AI can be created by other people in AI probably quicker. And so your time to defend the investment, you know, to build a moat is so short that the key thing with AI is distribution. So if you're one of our growth stage businesses, like a devoted or an Allidate, applying AI to your business can be awesome because you already have an installed base that you can then push it through and do things. That's why in healthcare, EMRs are in a pole position for AI because they have users and they can push it out.

14:06Bob Kocher: And so I think a lot about when I see AI, what is the way to distribute this thing? So you get it into the workflow and people habituated to it, then you're not going to then have people just switch off of it. I think AI by itself is really hard to invest in because it's so easy to copy. I feel like a lot of companies that I saw a year ago, six months later are free features of OpenAI or Anthropic. And so they just evaporate these companies. I like things that are full stack. So they are doing the care and they can use AI to make the care better. I like things that bear risk. I like things that have gone through thorny, long, hard contracting cycles with payers.

14:41Bob Kocher: because those end up being the things that are defensible or things that have distribution already. But AI by itself is wonderful, again, but I think it's more wonderful for our growth stage business to add 10 points of gross margin to their business and new features than to be a startup today trying to compete in AI.

14:59Unity Stoakes: Yeah. I mean, it's just amazing to watch how quickly, say, OpenAI is improving with their new health product.

15:06Bob Kocher: I mean, their new health product without training appears to me to be as good as any decision support tool I've ever seen. So that's a problem if you built a predictor of sepsis. Yeah.

15:18Unity Stoakes: Gosh. So, you know, interesting time when things literally are changing by the week, as you're a founder, as you're pitching someone like you, as you're building. So take us inside your head when a founder pitches you and let's just, you know, assume they have a good idea, I guess. But in the first 10 or 15 minutes, what are you trying to figure out? What gets you leaning forward or vice versa? When do you check out? How does that?

15:49Bob Kocher: I try to listen and learn. For me, the most exciting part of meeting a founder, I guess are two things. One is why do they quit their day job and start this company? I mean, there's always a really fascinating story because every founder I met was doing something different before they started their business. And it is a scary, giant leap of faith to go start a company. And it's like having a child, you have to bring it to life and bring people to it. It's like the hardest job that exists, I think, is starting a company because we have to make something out of nothing. So understanding why, like, why was this the problem that they quit their proper job and spent their life on is like really interesting to me.

16:26Second thing is, and probably where I spend

16:28Bob Kocher: the most time is how big is the impact? How big is the effect size of the invention that they have? I think a early stage business, you have to be able to see from Jupiter that is a better approach that emits photons of value to people who are using it. It can't be like a p-value of 0.05. It has to be like, oh, my God, it's obviously better. And so I'm focusing on like how much better is the innovation than what people are doing today? And then how many dollars does that generate for the person who's going to be paying for it? And it better be big. But that's sort of number one. companies that I pass on quickly are ones where the effect size isn't that big.

17:04Bob Kocher: So they never get over the inertia of like whatever the status quo is. So effect size, number one, after I think the effect size is big and awesome. Then I go into like all the versions of the questions you just asked around. What's defensible about this? How do you find it being copied? Is there a network effect? Like why can't, you know, once you show people that this is awesome, why can't every MBA student just make one like then what? but the first thing is really effect size and what's cool about or as it started is sometimes people have really imaginative approaches to problems that are hard and they have really big effect sizes i remember when i met sami inkan and the verda founder and ceo he was like bob did you know that we can cure type 2 diabetes and i'm like i didn't know that like i wasn't taught that in medical school or residency i worked at jocelyn diabetes center like that's like the best diabetes center in the world they never talk about reversing diabetes and he's like well here's the data and I'm like, oh, it's a big effect size.

17:59Bob Kocher: And so you can't not like that impact. And then just spend all the rest of the time on, well, how do I build a business that works to do this thing? Because for some reason, we haven't chosen to do this thing as our clinical practice. But effect size is the thing that I get really excited about.

18:15Unity Stoakes: And is Benrock doing, what stages are you starting?

18:19Bob Kocher: We invest at company formation at the beginning. like we co-founded a company health and devoted health and seed validated at the beginning or the a or the b but like kind of the earliest stage out of the venture fund then we have a growth fund

18:33Unity Stoakes: that will invest later stage in the portfolio too and you know obviously things have evolved a lot from seed and a from when you started to to today what what do you think founders really need to have proven today that might have been different from years past where you really need to show up with more today?

18:58Bob Kocher: Well, certainly have more software built at the beginning, because when we were seeding things, when you and I were starting Sordicholitan, I joined Venrock, we were looking at screenshots on a PowerPoint of what our product might look like. And then they'd be like, you invest some money and give me a year and I'll build the thing that you looked at on this PowerPoint. Today, people can arrive with prototypes of the product. And ideally, they've had customers try and give it feedback. And so you're not limited so much by a very hard to build software product. And now the product roadmap is like whatever the customer wants, you can build it pretty quickly.

19:33Bob Kocher: We spent 10s of months of hours at Devoted building Orinoco, which is the name of the claim system and the operating system for Devoted. I heard that companies have built claim stacks now with like six people. Software is the biggest difference that you can have a lot more of it. It can be a lot more ready. And probably even at the seed, you have at least a very good prototype you can show. And that's why then you go up to

19:57Unity Stoakes: the real hard challenges of actually closing those big contracts that might take, might be

20:02Bob Kocher: impossible to close. To get the A done, you need to get somebody to pay you money for the software. I have some contracts that look like they're sticky. Or if you're selling to consumers that you have CAC, ARR, and churn data. So one can make an assertion, you know, a certain discern, like, is this a business that actually grows or is the churn going to kill it?

20:19Unity Stoakes: Yeah. Well, let's shift gears a little going back to prevention. And you spent so much of your career thinking about business models around prevention or policy changes to try to improve prevention. There's a lot of founders building or trying to build things in the prevention or longevity or continuous health businesses. But it's just not clear that the business models can work or what they are. So if someone in the room or watching were building a prevention company today, what business model would you tell them to pursue, I guess? Or how would you advise they think about this challenge?

21:05Bob Kocher: I think it's the hardest problem because the incentives are most misaligned for prevention. I think the two business models that can work are direct to consumer, where you engage people because they find it entertaining or they get a great deal of joy from prevention, or in Medicare Advantage where members typically stay with their Medicare Advantage plan for five to seven years instead of you have some time to do it. But that's less prevention. it's more I'd say like making the diseases you have not get worse it's no less prevention but those are the two populations or the two different businesses I've seen you need to have somebody in your population for prevention for several years for the prevention to pay back so that's why Medicaid manage is where I think of because those people stay longer otherwise if you're doing it for people under 65 I think it's DTC and if it's DTC it needs to either give people really great joy from the experience or a lot of entertainment or, you know, be super engaging because otherwise your churn is going to be a problem in that business.

22:04Unity Stoakes: Like what Aura is doing?

22:06Bob Kocher: I'm wearing Aura. You are too. I find it entertaining and it gives me joy. And so I'm not buying it for prevention. I'm buying it as a consumer entertainment device, basically. I think most people are. I don't think of Aura as a healthcare provider. Like they can try to pitch me on telemedicine. I'm not going to buy it because I'm, you know, intrigued by my data. and so that's why I pay for it. But to me, it's the entertainment product. I do not think insurance will pay for RoraData or that doctors are going to have a desire to have an API to it. I think Function Health and, you know, other consumer health businesses are going to be DTC because they're not going to have an ROI for an insurance company to pay for these things.

22:46Bob Kocher: And so they're going to be entertainment businesses or, you know, business for the curious. But they're going to all then struggle with CAC and churn. Yeah.

22:56Unity Stoakes: Interesting. Well, my dream is we got to crack the code on this.

23:03Bob Kocher: I'm an author on a National Academy of Medicine paper coming out for ways to better line incentives. And one of the centerpieces of what we're recommending is multi-year insurance products. So if we had five-year insurance products in the commercial segment, we'd have five years at least to show payback for prevention. So that would make a bunch of things possible that probably aren't possible today. So that would make things like weight loss cost effective, mental health care access and screening and better treatment. Certainly nutrition counseling would become much more cost effective, like would make more sense.

23:37Bob Kocher: And so it'll unlock a lot of business models. Sleep would be actually something that we could pay money to improve. And so I'm excited about the idea. Unfortunately, it's an idea still. We need legislative action to do it. It's not possible to do it today. But we make a compelling argument that it'd be good for Americans to long term make health care. We think the cost grow less.

23:58Unity Stoakes: We're going to shift gears to audience questions in just one second. So if you've got a question, there's some good ones coming in. Please post now. But while people are posting more questions, just for the founders who want Venrock as an investor, a lot of them do. What's the best way to get on your radar? What makes you want to even look at their email or is it coming in through someone else or how do you get on your radar?

24:22Bob Kocher: Believe it or not, I read every email I get and I respond to most of them. But it's always nice if you come referred to a friend. That's nice, too. I think the most important thing is that we begin getting to know you before you're doing your fundraising, because when somebody invests in your business, they're investing in that business until that business is done. And that's like a marriage. I mean, that's like a relationship that you're entering into. And if I'm only meeting you for the first time when you're meeting 25 other people in a one week long sprint of go pitch investors, it's probably not enough time for you and I to decide if we should spend the next 15 years together, like talking all the time.

24:59Bob Kocher: And so for me, most of the investments I make, I've known the founder for somewhere between six months in years. And we've spent time talking about their business. And for me, it's the dialogue around the business, why they started it, the effect size, the business model, what we're learning along the way. What could we do better or different? Like that's actually how you decide if you want to work with somebody. Because if I invest in your business, I'm going on your board and we're going to be on a Slack channel and a text chain and talking at least once a week until you're done. And you better like me, I better like you.

25:32Bob Kocher: We better have a way of working together. And so if you're just looking for money, which is sometimes what you want, then I'm not a good fit for you. But also, like, I only make a couple of investments a year and I spend much time on them. And so I have to make sure that it's going to be a mutually good fit. And that's, I think, that'd be advice to give any founder is that the time, really, your best time for fundraising is not when you're fundraising. It's getting to know the people before so that when you actually fundraise, now's your opportunity to invest in the business. And then we've done the work.

26:02Bob Kocher: We know you. We understand the business. And then it doesn't take long to figure that out.

26:06Unity Stoakes: Great, great wisdom. Thank you. And thanks to all the health transformers for your great questions. And most of all, thank you, Bob Kocher. And really just what you and Venrock have done in building so many extraordinary companies and supporting and investing in so many great founders. In fact, one of our next Fireside Chats is with Todd and Ed Park. So we're excited about that. But just want to thank you again, Bob, for sharing your wisdom and spending the time with us. Thank you so much.

26:36Bob Kocher: Thank you for having me. And I thank you so much for the work that you do and all the wonderful businesses that you're building. So I can't wait to learn more about them. Have a great day. Thank you.

26:43Unity Stoakes: Everyone, see you soon. Bye. And finally, an invitation to all of the extraordinary health transformers, whether you are a startup, someone building the future of health, or whether you are an investor, funder, or buyer of the solutions out there in the marketplace, this is an invitation to you to join the Startup Health member community. Learn more at startuphealth.com. Look forward to hopefully seeing you soon. Thank you.