The Next AI Advantage in Healthcare Starts With Clean Data | with Pablo Perea at GoHub Ventures, BOAC #60

0:00When we're doing PC, there's no technological due diligence. It's more of an experienced type of due diligence in just how open-minded they are or how well aware they are from what's happening in the market. What we do at GoHub, we're an early stage fund. 80 % of the fund is going to be focused on digital health companies. One of them has been, for example, this AI layer or interpretation layer on their own health data from wearables and so on. So in our case, now that's not part of the thesis because we realized either we were late to the party or now it's non-defensible. So that's why I think on a longer term, purely like horizontal plays is not the right way to do.

0:38I might be mistaken, but that's my opinion. And we feel like this AI-enabled services is the perfect play. Some would argue it's not as scalable, but I find that now with how AI is evolving like different tools that you can get, I think the opportunity to invest in those type of companies is either now or probably 12 months ago. If you want to be on the earlier side.

1:07Hi, everyone. This is Jared, co-founder at Coherent Healthcare. We're back for another episode of the Business of a Clinic. This is an exciting one. I'm excited about this one because it's going to have a slightly different shape, slightly different feel. previously and usually we have healthcare operators, executive operators, people investing in healthcare in terms of acquiring clinics, building groups, entrepreneur clinicians, people who are really front lines in seeing patients. And so much of the conversation is about technology and AI and efficiency. And there's a completely different side of that conversation that we never really get to hear about.

1:44And it's the story of how do you get innovation from someone's brain into the market. And what does it take to actually do that? And oftentimes it takes capital. And so today I'm very excited to have Pablo with us, who is a partner at GoHub Ventures, which is a Spanish venture capital fund specializing in B2B, but also largely in healthcare and in digital health. So Pablo, thank you so much for joining us. Very excited to have you here. We've known each other for about a year now. Can you give folks a sort of two-minute version of you. What do you do? What do you do exactly? What does a day in the life look like?

2:18So, yeah, thank you for having us. What do we do at GoHub? We're an early-stage fund. We're investing from Spain, but basically it's about the world. Basically it's mainly on Europe. And we're mainly looking at a digital health company. So we've decided that this is a key sector both on an opportunity basis, but also on just like a personal level, our fund, our LPs, they really want to invest in the sector in this innovation. And so probably like three years ago, we decided to go all in almost like 80 % of the fund is going to be focused on digital health companies. And it has been a learning process.

2:59Definitely. The first year was tough. We saw many companies in the space. And since then, we've narrowed the thesis quite a bit. And the thesis is quite flexible, but you have to stick to what you understand as an opportunity and where you want to invest. And what we usually do at GoHub, we have a defined thesis that every six months to a year, we see what are the other sectors that could come into the thesis or what are the ones that we probably see that the opportunity probably is there. But we decided deliberately that maybe it's not the main thing or the proper timing for us to go into those companies.

3:37And so we're ever changing the thesis that we have. And yeah, now investing between six to 10 companies on a yearly basis. Spain, UK, Germany, Switzerland are the key markets for us, but we're fully flexible on that side. And we're mainly doing pre-seed, seed, and some, should we say, opportunistically. Narrowing focus on health. Yep. Pre-seed, seed. So for those less familiar with the venture cycle, early days. Early days, things sort of coming into market and then I'd see more maturity starting to see more adoption. Series A obviously starting to grow and see that moving forward. When you say it's been a learning curve, I think it's been a learning curve for everybody in health care for the last year and a half.

4:21What is the, I guess, the tip of the spear for you now? You've sort of seen the first wave of tech happen in the last 12, 18 months. where have you landed, I guess, personally or at a fund level in terms of the things you believe about healthcare X technology more generally? There's like several examples. One of them has been, for example, this AI layer or interpretation layer on your own health data from wearables and so on. This is something that probably like two years ago we started like looking at startups there and now like everyone's doing their own or there's like plenty of companies in this space some of them are really well funded and doing really well and some others like they just decided to stay in the middle and and then they usually like you see plays like over our whoop and they they have like their own thing and it's doing really really well so in our case now that's not part of the thesis because we realized well either we were late to the party or now it's non-defensible.

5:27So that's part of the learning curve that we've had in the last years. And now, for example, last year, we do on the thesis level. Last year, we had this market deep dive and see where's the opportunity. And we decided that in digital health, there were like three or four main areas for us. So one was AI safety layer. So anything related with security of implementing AI in health, both of the public and private side, mainly in private. That's another learning curve that we had. That's one sector that we're doing also on what's the common denominator on AI operating efficiency within private clinics.

6:06That's the core sector that we're looking at now. And then also tools that are building on the payer side, so the insurance side of the business, especially those that are reimbursing technologies. This is the three sectors that we identified last year on top of obviously AI and data, which is the, I think, common stuff that we look at. And then throughout the year, you start seeing other types of companies in subsectors that you probably didn't recognize that in the first moment, things like hormone tracking. We've seen a lot of hormone tracking this year. In the pharma side, we're seeing lab automation stuff that at the very beginning, we didn't really see as an opportunity.

6:47And now those sectors, for example, some of them we've decided not to continue looking at them. For example, in the hormone tracking stuff, there's really well-funded players that we've seen those companies as well since the first couple of years. And now either they're super well-funded or they're still in the technological development stage. And our Our understanding is probably the technology is not in the right moment. So for us, we've deliberately decided to stop for now in order to look for those purely hormone tracking companies. Probably next year, we will reconnect on those. But for now, I think it's a learning curve as well.

7:31And then sectors that we've identified now throughout the year that probably at the very beginning wasn't or they weren't a key sector for us was probably all of the fintech side of health has always been interesting. But I think now with like compressing margins and everything, maybe it's a little bit better timing for us. And we're seeing the safety layer as a key thing, which for now, we're looking a lot into compliance side, which is not security, but it's kind of in the sector and slowly incorporating these other types of technology, for example, like regulated software or new clinical pathways like mental health or sleep technology, like really hot markets.

8:20There's not a winner in any of those. You can argue that Headspace, for example, in mental health is related, but there are a lot of companies doing voice AI agents that recognize if you have a pathology or not. And that's what we're looking in mental health. We're seeing the same in sleep, like sleep apnea, like how to treat sleep apnea. Like we're ages away from, like we should have already found something that works better. So yeah, these are sectors that we identified throughout the year. And now it's part of the thesis for next year. And in six months we will reevaluate and see if it makes sense.

8:55Probably we've invested in some of these companies. And if that happens, we can still look at them. but probably it's not the core aspect because you're already present in that specific market. So that's really helpful and I think gives us a broad picture of the different things that you're seeing and obviously you're seeing or you're trying to see these things 36 months before they hit high street, right? Because a bit of crystal ball, what does the future hold? Obviously for us at Coherent and a lot of the audience and the folks we regularly speak to, it's really on that clinic efficiency story.

9:35I'd be curious to understand the journey you've been on as an investor in terms of clinic efficiency. What has you excited? What has you less excited? And I think similarly, if you imagine the customer here, a customer being a clinic decision maker or someone who's actually a healthcare operator, private hospital, clinic group, whatever it might be, at least in our experience and having previously run clinics so much of the challenge in bringing healthcare innovation to market is regulatory safety compliance consistency etc those are hard things to get right but also it's just mindset and inertia and healthcare is historically a slightly slower moving animal it adopts more slowly things change more slowly the system needs a bit more safety built.

10:26So I guess to be curious to understand from an investor perspective, when you think about innovation in clinic efficiency or however you would define that thesis, what's going on in that? Several questions within one. My understanding first, probably on the innovation side, so on the startup side, What we started looking at probably two years ago, honestly, I think it's the clearest example or the most prepared example in innovation within health to incorporate technologies in the clinical, like in the private clinical space. So for us, it's been something that we've been looking at several years already.

11:09And probably at the very beginning, we were looking more on AI, like purely AI companies that are doing either a verticalized approach to a specific type of clinic, like dental clinics or whatever. or what we've seen probably more recently is just like pure horizontal type companies that are all AI super homogeneous amongst themselves and they serve to several types of practices right so dental, MSK like several others and what we we've come to realize after speaking to many of them well first of all there's many of them so it's definitely a sector that there's a need but there's also a lot of providers and what we're seeing with those is defensibility is not key it's quite homogeneous and if you're looking to invest in or you're looking to penetrate into private clinics that are in different sectors probably like 30 40 percent of what they do is similar, but the rest of that is not similar.

12:14And that goes into not only the, obviously, like the intake and so on, it could be somewhat similar, like patient intake. But if you go more on the clinical side, nothing to do between one another. And so we are seeing that this horizontal, like purely horizontal place, which now when you find these companies, they seem interesting probably they're not they can tap and distribute into this type of clinics but I feel like it's not a long-term play it's probably something that can probably tick the box and the very beginning but suddenly you see that there are like hidden costs to that like you're probably need a we've spoken about this before but probably there's someone that is dedicated to just like how that works and you're just incrementing the costs of the overall business even though you're looking for the efficiencies I told.

13:07I think in that case it's not only the responsibility it's on the clinic side probably you don't get the ROI that you're looking for. So that's why I think on a longer term purely like horizontal place it's not the right way to do. I might be mistake it but that's my my opinion and then counter to that either like you have those purely verticalized plays which are good but if you go really verticalized it's it's tricky then to go and open like different geographies or different like sub type of clinics within the same vertical and I think that could create kind of like a Frankenstein type of product and And I think that's a little bit even better than a horizontal place.

13:53Then you can argue there's this hybrid place, which is kind of horizontal, kind of vertical. And it's on a modular basis that that works as well. But there's several players. And then the other side is purely like AI-enabled services place. So companies that are using AI and somewhat similar kind of architecture, and they're using that in order to prove purely the ROI that the clinic is looking for. And that in my view, it ticks both boxes. It ticks the boxes of, I want AI in my business to create efficiencies, and then is it creating efficiencies, which is the thing that you're actually looking for.

14:32And we feel like this AI-enabled services is the perfect play. Some would argue it's not as scalable, but I find that now with how AI is evolving like different tools that you can get, although people are tech aware and they are getting comfortable using this i think is increasingly like getting faster adoption than one year ago definitely and probably this is the type of companies that if you can do this a enabled service and you can really get into the clinic and like deep down into the clinic that's defensible you get to know the business you to know the vertical. And I think that's like this snowball effect in terms of ROI, like as deep as you can get, the more ROI that you can get.

15:18And that on the clinic level, I think it's clear that if you have this long-term view of, I want to be efficient and I want to have clearer ROI in my business, you have to not take this leap of faith, but really understand that there are this type of plays that are kind of like a partner for the clinic. And if you are willing to commit on that side, probably there's no hidden costs. It's only like increasing margins in that case. So that's my view of how we see now AI implementing into pallet clinics. And it's the same type of play in the B2B sales side. There's a lot of AI cold agents, like cold calling agents.

16:01And there There are places that are service cold calling agents. And I feel like that's on a personal level as well. It just converts easier. So for me, like distribution, differentiation, and then the ROI on the clinic or the customer, it works. It's interesting. We spoke recently with a dental group leader actually out in the US who is very tech forward. And I think he didn't use investor language, so to speak, but I think he was basically saying very similar which is that he had i think they were 70 or 80 sites and initially they tried ai receptionist for all health care found it was too broad and too much was falling through and you know you saved five minutes here you spent 25 minutes here so it was it was so horizontal that it didn't deliver on the promise and then i think they over corrected and went the other way and they said actually we're gonna bit we're gonna go and try and find a solution for this like super super tight periodontics whatever it was um and that was really good at delivering on the promise but it only delivered on the promise in a really specific orientation and then again you're having to catch the overflow um and i think that's one of the challenges that we've seen in ai only solutions so far is that you can't so you can't solve a hundred percent of the of the instance And so the question is, whether it's 50 % or 70%, 80%, whatever it is, I think what we've learned on the journey is that the data is such an important part of realizing the benefit of AI efficiency, that if you don't have hygiene and the ability to interpret the data, you sort of end up with spaghetti all over the place.

17:50I'm curious from an investor perspective where you are seeing AI work, maybe in the portfolio companies that you've invested in or things that you've seen more broadly. What are the sort of ingredients for AI and healthcare really working? What are the one or two things, like the key things that have to be true for it to work? So from what we're seeing in portfolio companies, one is on the internal side of the business, like development, like product development, like new products. We have a couple of companies that are not infraplaced, but they're kind of on the data homogenization of the companies, which the value add probably for the clinic is not as clear when you go and sell to them.

18:36but then ai has enabled them to create like super fast shop products or modules on top of that which is non-defensible but it's the only way to really like see clearer value from the point of the view of the of the of the client in this case that you're trying to sell to so ai has definitely helped them in selling the company selling the product and also selling uh or creating different types of product that increases the whole footprint of what they're aiming to do with the client. So that's one. And then on the other side is, I believe there are a lot of AI type of tools that help on the data cleansing side.

19:20So that's more probably on the client side, because there are a lot of clients that are aiming to integrate AI and they don't realize that within their own company, like everything's not as tidy as they think. So I think probably this is not helping directly our companies in the portfolio, but they're helping preparing the client in order to sell to them because some of them that they really have this long-term idea that, yeah, I want to integrate this type of technology within my practice. They have to make sure that when they integrate it, it's going to work and it's not going to be like spaghetti all over.

19:59And so this type of client that they have this long-term view and they either decided to do it on themselves or they hire some other company to do this data cleansing or preparing the structure for that is indirectly helping our own company. And honestly, like most of those clients, they're not well prepared just to integrate AI and suddenly everything's working. that's why also going back into the AI enabled services I think it's even I think it plays in favor of this type of phase because not everyone is super prepared probably like larger types of groups that have been doing this for several years already and they have their clear procedures and everything probably is a little bit well prepared versus like a smaller type of clinic that suddenly they've decided to do this change But honestly, the vast majority of clients that are trying to sell to, they need to have this preparation internally.

21:00And I think that AI is helping and it's helping us as investors as well. So definitely a lot of points, but I think the clearer side is the product development and sales is helping. It's interesting, you used the word preparing for this idea of preparation. and what we often find ourselves surprised by is some of the customers that we work with, they have already been down one failed experience, which obviously makes our conversation very hard. It's like you got burned. You got the scars all up and down your arms because you just lived through six months of, you know, all the mess that comes with it.

21:41And one of the things we always ask when that conversation presents is we'll say, that talk us through the first month of working with whatever you were just using before you went live. Before you started doing anything, what did you do? What was the onboarding like? What was the documentation like? Share with us whatever it is that you did then. A, it helps us pick up from not zero. And B, we get a better sense of where some of the gaps are. And one of the things that we're always very surprised by is there's nothing. because it's a self-service tool. You go in, you put your API key in, suddenly it talks to your booking system, but it doesn't know anything about who to book, when to book, how to book, where to book.

22:26SOPs and everything. Totally gone. And I think this has been our biggest learning as well. So this idea of preparation is that actually to get the value, you have to do a lot of upfront work. It's very hard work and it's very unsexy work. Writing everything down is not sexy. No one likes writing everything down. Now you can talk everything down, but it's really unsexy. I'm curious, when you think about due diligence, obviously the way you do due diligence as an investor is different than the way a customer would do due diligence, but there's probably some overlap in the sense that the things you're looking at as an investor are probably things that a customer should look at as well.

23:11And preparation or onboarding, as we would call it, is to me like one of those things, like how deep have they really gone to understand that problem? One of the things I always ask clinic leaders is what's your AI or tech decision? How do you make decisions about technology? How do you do due diligence on your suppliers? As an investor, putting aside, you know, company, whatever, how do you do due diligence? And, you know, maybe to those listening, what can a powerful possibly borrow? Because you look at tech all day and this is literally your job. Whereas a clinic owner or clinic leader, their job is not to look at tech all day.

23:46It's to serve patients. So surely there are some things, some patterns. Yeah, I was going to just answer patterns. Patterns both on, so when we're seeing companies every single day, right? And even if it's companies that are not focusing on the same type of problem or same type of market even, they probably have a common type of roadmap. So whenever you're seeing another company, like you can think back on the previous deals that you've seen or that you have in standby basically for the next year. And you compare this type of roadmaps. And when you see similarities and you honestly, like from what you see, just learning about technological stuff and like news and so on, when you have your own like built decision of what is probably the best path on the technological side, and you see those patterns, it helps you create conviction that probably their roadmap is what it's going to be or where everyone's heading at least.

24:48Probably it's the complete opposite and the ones doing it. The other thing is going to be the ones that are working. But I think if you think on the bigger picture, everyone tends to go on what is more efficient or what could work best. So that's one, which is just pattern level. But then also part of our investment analysis is also speaking with clients, like nonstop. And we speak with clients either from the company that we're trying to analyze or trying to invest. that we have like this kind of referral and they see like probably they're working with this company for like a year or a couple months.

25:26And we have this comparison of what is a point of view of working in the short term versus a long term and how it works and how the company has been evolving with this client and so on. So that's one check that we do, which is more on the relationship level rather than like purely technological one. But then what we also do, And there's also some kind of pattern here when you speak with different types of customers across different types of companies that we're investing and then companies that we see in our own personal lives or our own ecosystem, like for us in Spain or when we're traveling and we meet other type of partners.

26:06You start seeing what is the need of the market. You start understanding, you listen to a podcast and you understand what is the common need in a specific market. And so you have to really start making ties into, you know, what is the roadmap? What is the need now? Then you see how technology is evolving and you see, okay, the need could probably evolve into this. And so from the investment side, you're looking always forward, like forward looking, like you're investing for the best of the company, like how big can they get or how can they become a larger company? and so you have to think at least in 12 months time ideally like 24 months or 36 months time and for that to happen like you have to find the common like crossing uh point between like technology client need market just readiness and everything and then also like there's a duty that is purely technical you have your own advisors you have um advisors from the let's say market level, also on the technical level.

27:09And you always leverage those type of contacts in order to analyze a company because they're probably they've been dedicating themselves, like their own lives into that specific problem. And they know what works and what doesn't. And just a clear, let's say, understanding or an opinion of, I honestly think they're on the right path. That just helps. It helps a lot in the analysis. and sometimes like when it's in the pre-seat stages which is probably product is just an idea or it's just like a initial version of what it could become that's more on a conviction play on the team like you have to really after speaking with the team getting to know them also on a personal level on an experience level like previous experience that they've had when you see that and they can genuinely create your like they're convincing you that they are the ones that know how this is going to evolve based on experience.

28:06Sometimes you just take this leap of faith in pre-seed, which is basically what it is, and it's a conviction play. Just thinking that this type of founders, they really know what's going to be evolving. And AI is happening a lot, like people coming from academia and so on, in health, obviously, as well. And so when we're doing pre-seed, there's no technological due diligence. is more of an experienced type of due diligence in just how open-minded they are or how well aware are they from what's happening in the market as a general idea. Makes sense. I think one of the things I'm taking away for clinic leadership in particular is obviously customer references.

28:45That's not a... Many of our clinics will ask for customer references, so that's not new. Would be good. Yeah, it would be very good. But I think what is interesting is this idea of technical advisor because you are deep on technology and you as a fund have still made a decision to find people who are even deeper. And many of the clinic leaders that we speak with without intending in any way to disrespect them, there is sometimes a technical gap. I mean, when I was running clinic, there was definitely a technical gap as well. There's still a technical gap depending on what we're looking at. And the idea that actually having a sounding board that can help you figure out if directionally it makes sense or hey has something that they've said just did they say something that is basically like stepping on a landmine and i think we have come across a few healthcare organizations where that know-how does not exist even at the c-suite level and the ones i think who have really overcome that challenge they are almost creating their own micro tech advisory board through their own networks completely out of healthcare you know maybe they went to school with someone maybe it's the friend of their husband of their wife of their whatever and we've had a few conversations now with clinics where we're in you know in discussion with them there are emails going back and forth and the the outcome of the fifth email is, hey, would you mind just speaking to so-and-so?

30:18They've just, they help me, they work with us and they have a few questions. And it's so refreshing to see group leaders trying to close that gap for themselves. I think in a similar way for you, maybe the gap is a bit different, but when you hit something that you're not totally sure on, actually it's good. Like go find that expertise, I think helps. Final question, wanting to respect the time you've already given us. you said a lot of your job is trying to predict the future and figure out people who are generally moving in that direction and hope that things sort of match up and marry up at that time.

30:56If we were to sit down in 12 months from right now, where are you most confident today in terms of what the future? I'm sure you're not confident about a lot of stuff, but what is the thing you're feeling right now that feels like it's got staying power of all the stuff that you see on a regular basis? I'd say on the data side and the data readiness, I think that's like, if you want to be data ready, like you can do it as of now. So my idea is probably now first early adopters are already in that stage. And those that are not now, they will be in 12 months if they are like forward looking type of company.

31:38So data readiness and just willingness to adopt AI, that's for me, for sure. And probably now there are technologies that are a little bit earlier because they cannot sell if the client is not ready. So for us, I think there are going to be sectors that are going to be unlocking in 12 months time in order to invest. So that's one. And then the other one is more on the overall efficiency of the company, like both on the, let's say, clinical efficiency and operating efficiency. Probably more on the operating efficiency because it's less regulated and it's easier to adopt. So things like patient intake and then triage and so on, I generally think that's a ripe market now.

32:26There are really cool companies growing massively super fast in this type of like niche solutions and my idea is that this whole let's say patient facing operation which is on the operating side of the clinic that's going to be probably not solved like there's going to still be a plenty of market but i think the opportunity to invest in those type of companies is either now or probably like 12 months ago if you want to be on the earlier side probably in 12 months time these companies are already scaling up which is now the series a and above the stages um so that's probably the two aspects then the clinical side i feel like it will come down to the type of um procedures that they're doing so things like probably um so heart related diseases there are a lot of investments to being done uh over there and now it's everything subject to regulation like you know to get the the FDA or whatever type of regulation and that obviously takes more than 12 months so it depends within the pathology that you're looking at if it's more or less ready or not that's why I think in the clinical side there's many like sub areas and some of them will be a little bit more prepared now whereas for example sleep technology which is something that we're looking at we believe the the opportunity is there for someone to get into that because there's not a long-term or cure winner in that category as of now.

33:59Same with mental health is happening quite a bit. So I think that with clinical, it's a little bit challenging versus the operating side for sure. But both operating and just data readiness for me is a no-brainer. Anyone that's thinking on a longer term, they're going to be really well-prepared and those are already testing whether if it's like in-house type of tools that they're trying to build their own uh like tech stack and they're bringing advisors or just hiring people that are super ai ready and they they're investing their own time and obviously the resources into that or um you're just trying third-party companies and sometimes the that would be really good and And sometimes, obviously, there will be scars about that.

34:46But if they are looking on a forward website, sometimes having a scar because you've used a, let's say, horizontal plate, going back to the first question, probably you realize that probably instead of horizontal plate, it's more of a service or vertical type of plate. So that would then help and accelerate the distribution of these other type of companies. So that's why I feel like in your case, which is more AI-enabled services and more on a deeper approach and demonstrating clear ROI on the clinics, that could pay off both on just market readiness on technology, but then also on companies that probably they've tested their own in-house tools or third-party tools, non-defensible tools, whatever, and they've realized that there are hidden costs, the ROI is not as clear or is not as big as they thought it would have been.

35:44So all of the investment that they've done before in terms of resources time and everything doesn't really pay off. So I think it's going to be a really cool market situation next year in terms of how these same technologies, but on a different, let's say, approach are going to be working. also how our clinic is going to adopt that. But for me, it's a no-brainer that in the private side, especially it's going to work really, really well. Cool. Pablo, thank you so much for joining. Folks, thank you for joining. It's really interesting to have the perspective different from the conversation we usually have.

36:23And folks, if you're interested in speaking to Pablo, you can find him as part of the GoHub Ventures team based in Spain. And I'm sure you'll enjoy reading some of their stuff about the future of healthcare, future of digital health as well. I really appreciate your time. Thank you for having me. Pleasure.