Sword Health CEO Virgílio (“V”) Bento on Vital Signs

Sword's New AI Solution and Decisions That Shaped the Company

A few weeks ago, Sword Health announced a new funding round at a $3B valuation and introduced Phoenix, their new AI care specialist. We sat down with Virgílio (“V”) Bento, the Founder and CEO of Sword, to learn more about Phoenix, the evolution and key decisions behind Sword’s success, keys to employer sales and more. You can check out the full discussion on Spotify and Apple, or read the highlights below.

Highlight 1: An overview of Phoenix

V explains that the number one problem in healthcare is patient engagement, which is why Sword aimed to make its AI solution as human-like as possible, enabling it to converse with and adapt to each patient. During a session, it can measure the patient’s movements, identify errors, and suggest areas of improvement.

“Now the important piece here is that it serves as the co-pilot where what this means is that every single action needs to be approved by the PT or rejected or edited, so nothing goes out without the PT approving that. And that's very important because, well, I will definitely not feel comfortable to have an AI system working, being live without guardrails. So the care specialist works as the checks and balances of the AI system and also expands the AI system in terms of ability. And then of course Phoenix then really scales and provides leverage to the PT.”

Highlight 2: Structuring an organization for growth

When Sword plans to expand into another condition, it hires a strong generalist and provides them with $1 million to put together a team and reach product-market fit (PMF) in 12 months.

“We want to replicate that scarcity mindset, and then after 12 months, we discuss -- was there product market fit or not. If there was, you do your series A, which means that you get more money, you get to accelerate faster. If it doesn't, then we kill the solution. If you go to product market fit before the 12 months, you are able to raise your internal series A faster and then you expand.”

Highlight 3: Breaking in & building trust with customers

When selling to employers and health plans, V highlights that there are only two types of solutions that can scale: those that allow customers to save money, and those that enable them to make more money.

“You can do two things and we are doing both things. One is talk the talk and walk the walk, so contractually guarantee to clients that you're going to save them money. So, if you're going to say, ‘Hey, I'm going to save you 50% of the cost, let's put that in the agreement, and after 12 months you assess that, and if you did not, you pay me back money.’ The second way to do it is we have independent third-party companies evaluating your ROI. Because that provides some assurance to the market, and I think in the current environment I think you need to do both, not just one.”

Highlight 4: An early, game-changing decision

When Sword first started, their vision was to develop technology for PT teams to use in their clinics. But, when they began running pilots, a lot of clinicians complained. These results raised a concern: if our product wasn’t working well in controlled environments such as clinics, then how would it work in uncontrolled environments such as homes? However, V recognized that offering the solution direct to patients at home would work.

“And the reason why it was successful right from the gate is because it's not like the system was perfect, but even with all the imperfections and the rough, rough edges, it was much, much better than going to a clinic, asking your wife to get you there, find a parking spot, and then doing everything back home and then waiting in the waiting room for 20 minutes; even if the system wasn't perfect, it was 10x better than the alternative. And so right from that moment we understood, oh, okay, so we are not going to be a company that provides technology to existing teams. We are going to be a provider of care using AI.”

Here’s a full transcript of the episode:

---

This is Vital Signs, a podcast on cutting edge trends in health tech and the people shaping them. And today we have a guest we've wanted on for a long time in V, Founder and CEO of Sword, the company focused on improving pain management with an initial focus in the MSK space. They've raised over $300 million, most recently at a $3 billion valuation. And as V says on his LinkedIn, I think you're only 5% of the way there with what you want to do. An incredibly exciting company. Super ambitious V, thanks so much for coming on the podcast.

Thank you for having me. I really love what you guys do, so it's a pleasure to be here.

To kick things off, I think you guys have had such a fascinating journey. As I understand it, the company started in 2015 out of Portugal where there’s not traditionally been a ton of digital health companies, and I think it took you a while to kind of find those first investors that really believed in the company. Now looking back at the last five years, it feels like you've just been on this rocket ship of success, but I think that whole journey would be super interesting to our listeners if you contextualize the history and evolution of the company.

Yeah, it's definitely not the typical journey, that's for sure. So I started developing the technology behind Sword during my PhD in electrical engineering. And I decided to focus on that because I've experienced firsthand the challenges that families face when they have to provide care and recover their loved ones. And then I did my PhD, we had clinical validation, we solved several open problems, technical open problems in the field, and then we decided, okay, this needs to be a real solution. So, we need to build a company as more a means to an end to bring this solution to market at a global scale. But the thing was that these were a couple of guys from research, from academia, zero business knowledge, and as you pointed out, starting in Portugal, which is not exactly a vibrant ecosystem where you can find mentors to tell you, “Hey, this is how you raise money, this is how you build a startup.”

So it took us some time to access that level of knowledge. And the first five years, actually the first four years was really, was time to survive with small investments. I always tell people when they talk about fundraising that our first seed investment was $150,000, right? Followed by a second seed investment the next year of $250,000, followed by a third seed investment in our third year of $400,000. And our luck was that we got a European research grant of $1 million to accelerate the development. So it's a little bit different from 2021 where we raised our series B, series C and series D in one year and in November in two weeks we raised 250 million in just two weeks. So we've been definitely in both ends of the spectrum when it comes to fundraising, but it took us four years and then we were able to get a little bit more funding from some other seed investors in Europe.

And then when we went to raise our series A, we thought, well, we went to raise from the top tier VCs in Silicon Valley, but of course we have no way to access them. We have no network. And so let's start with some tier two investors in Europe so that we get some traction, some validation that this is successful and then go after those tier one investors in our series B. Funny enough, in six months trying to get that round done in Europe, we got almost 50 nos and it's not just a no, it's that kind of no, that doesn't go past the first meeting.

And so then late into the night I sent a cold email, basically a Hail Mary email to Khosla Ventures saying -- Hey look, you don't know us, but this is what we are, we attached our papers, our clinical validation, our results, we showcase our technology in an email. And again, surprisingly enough they replied and after four weeks we raised our series A from Khosla Ventures and that was surprising to say the least. And then of course some of those European investors when we announced the round, came back first asking us how we were able to pull that off and second if they could jump on board.

Actually, I just sent one cold email to Khosla because they were contrarian enough to bet on a couple of random guys from Portugal, so it worked out well.

That's awesome. It's amazing. And I think what would be interesting is rewinding back to that time when you raised the series A from Khosla and thinking about how the markets developed, if I remember how things were at the time, it felt like Hinge was the first to market and then it also felt like there was, I mean at one point if someone had made a market map, I'm sure they would've had 15 logos, the payers were trying stuff, there was a bunch of other people. You've obviously navigated this market incredibly well. How has the market evolved and what do you think you guys have done right to get from where you were series A to now you guys this week announcing your $3 billion valuation?

Yeah, so I think it's fair to say that very few folks were bullish on our success when we started. I still vividly remember a board meeting that I had on February, 2020. So basically we raised the series in 2019 and then we launched in January, 2020 and I remember we just launched in January, 2020, this was February, and I was meeting with our board and it was clear that our board members were again trying to understand how are we going to compete against this direct competitor that has been three years in the market, massive funding, lots of logos, lots of partnerships, and that in healthcare, no one really gets fired by choosing IBM, right? So how are we going to compete against this competitor when we don't have access to the same level of funding? We have zero clients in the US, zero proof points, we have nothing. And so the typical sign in the startup world is that distribution wins against product, even more so in healthcare.

But surprisingly we saw the opposite because again, and I don't want to talk too much about the direct competitor because they are not in the room, that's not our style, but the time that we spent building the technology, building the product, and really building a superior solution, it was very clear to the market in 2020. And that's why right out of the gate we started competing and in the first deal where we competed and it was a $1.5 million annual deal, we've won that deal. And then we started winning more deals and then in December we won a Fortune 50 company against that direct competitor. And then it's a snowball and you create a snowball effect, which is what's this new company? And it was not because we were better at marketing, honestly, it was not because we were better commercially, it was because we spent four years really building a superior product that was very vividly clear the differentiation to our clients. And honestly, I think that is the secret of our success is that we are a product centric company. That's what we do well. We do technology and product well, and I think in this market, especially in the digital health space, that's different. And I think clients notice that.

Totally. I mean I feel like at this point you're like a guru of selling to employers. I'm sure people come to you for advice all the time. Obviously in the beginning until you have those reference customers, it's challenging. What have you learned over the years about how to do that and what kind of advice do you give to other entrepreneurs who come to you to ask?

So I think it's quite important for you to think from first principles and to think about what are the reasons why employers or health plans choose a solution. And for me, I think there are, with the exception of mental health, which is a specific space in itself, only two solutions get to scale in healthcare. Either you allow someone to make more money or you allow them to save more money. And so our value proposition was dependent on how much we were improving quality of care, how much better patient experience and member experience we are delivering by enabling members to recover at home. It's really about saving money to clients and contractually guaranteeing that. So we understood that that was a massive value proposition and we went with that. And so it's really about understanding what's your unique thing is. Then understanding also, and this is a piece of advice that I give to young founders or young CEOs, is how you break through the noise.

Because if you talk to our clients, one thing they will quickly complain about and the rightly so is look, every single week I have 50 solutions in my mailbox box trying to get the call with me, trying to get my attention. And so how do you break through that noise, and there are several ways to break through that noise, but you need to find a way and otherwise you'll be one in one hundred solutions trying to get in and independently of how good you are.

How'd you guys do it out of curiosity or one tip you guys did?

It's not one silver bullet. One way we've done it right from the start was surround ourselves with advisors from the industry that could basically introduce us to the right folks because they have that connectivity because we had the social capital to access people. So basically we surround ourselves with advisors that believed in our mission, in our vision, in our solution, and they did that work for us. And so that worked very, very well for us in the early days.

And then you mentioned the payer side. How did you think about the right time to go to payers directly, and then how are those conversations or sales processes similar or different for you?

So payers are a very instinct beast in the best way possible in the sense that I usually say that in healthcare you only have clear proof that you save money for clients if you are able to sell that to health plans because they will in the first meeting or in the second meeting, you will meet with 10 or 20 actuaries that will dig through the data and if you don't have clear financial outcomes, they will know. And so the best way to get to payers is start with their book of business, start with their employers, showcase value to their employers and then they will notice that and they will be the ones or you can trigger that, but then it's much easier based on that data for them to expand into the commercial risk segments, fully insured Medicare Advantage where now they are paying for it. So, the stakes are much, much higher, but it starts with the employees and that's what we've seen. And honestly, that's a journey that is hard to travel because you really need to show outcomes. You cannot just appear on the surface that you have outcomes, which again, in healthcare as with many other areas and industries, that's a thing, but when it comes to payer, they will dig deeper and find the truth.

I guess there's a good segue to talk a little bit about what employers are looking for from an actuarial soundness or metrics and stuff they care about. So there's that report that came out recently about the diabetes management companies and it's like net of fees, these didn't really save that much money. They mentioned they're moving to musculoskeletal next and you have all these, I think new third party, I would say independent validators. I'm curious what you think about one, the role of these third party independent validators versus companies doing it themselves and how that differs from maybe what employers are looking for when they're comparing these different companies in the MSK space.

So, I'll start by saying that that's even more important now than before. We had our company conference two weeks ago and I was talking on stage with Nico, he's a SVP of benefit, and he was telling me that the landscape of the benefits space five years ago was mainly about improving quality of care, improving wellbeing; in the last two, three years, it's all about mainly driving cost efficiencies. And so it's not just saying that you save money, every single company in the digital health space has a slide on their deck saying that they save money, but the way they get to that assessment, it's sometimes quite creative. And so a benefits leader doesn't have really the bandwidth or the knowledge because this is specialized knowledge to go through the data and assess, oh okay, you really have an ROI guarantee or really have an ROI, a clear truthful value proposition when it comes to cutting and saving money.

So in that regard you can do two things and we are doing both things. One is talk the talk and walk the walk so contractually guarantee to clients that you're going to save them money. So, if you're going to say, “Hey, I'm going to save you 50% of the cost, no problem, let's put that in that agreement and after 12 months you assess that and if you did not, you pay me back money.” That's one way to do it. The second way to do it is we have independent third-party companies evaluating your ROI, right? Because that provides some assurance to the market and I think in the current environment I think you need to do both, not just one. And if you do both, then I think you have a much more credible solution and in the long term that's how you win.

I think a lot of people who enter the employer space, one of the things they struggle with is we make the sale to the employer and now you have to actually get the employees to use your thing, right? Curious how you guys manage to crack that? What are some techniques or things you do once you get the employer contract, now you're like, okay, we got to get their employees to actually use it. I assume there's some bespoke stuff per employer and I assume there's some stuff that transfers from place to place.

Actually, we are heavily incentivized to do that because we get paid on utilization. So our model is not a PMPM where you get paid independently of utilization. We only get paid if we deliver value. So of course that creates a big internal focus to be good at it. And so it comes down to the typical marketing tactics to have awareness in the population. Mailers directly to the employee's home and emails to your inbox. Again, you can tie multiple other things, you can tie webinars, that's okay, you can tie lots of other stuff. It comes down to those two channels as the driver of the majority of our enrollment. And so then it's really then working with clients to get that level of marketing access and that in itself is always an important conversation to be had because of course clients need to select which solutions get more marketing access; otherwise you are spamming left and right your population, but it comes down to those two channels -- mailers, emails, those are the two most important channels by far to us.

One thing I definitely want to make sure we hit on is, so I think by the time this podcast is released, you guys will have announced this really exciting set of AI features under Phoenix, which I guess is, I understand it is basically kind of an AI care team sitting alongside the human care specialist and working side by side and I think it sounds like as you were building this, one of your key priorities was let's make sure when the AI care model is interacting with patients, it's kind of indistinguishable in some ways or just as good as the human side. I'd love to know what went into building that. I imagine you can't just take GPT-4 off the shelf and then throw it at your patients and hope it works. So how did you actually go about building that and what were some of the problems you had to overcome to get it to this level? Also, why is it called Phoenix? Did this rise from the ashes of something else?

Because that's what's recovering is all about -- rising from the ashes, getting back to your full self. So it's not that deep though.

Beautiful.

So first it's not using AI just for the sake of using AI because it's cool. Well first it's cool but it's actually solving a problem. And if you look at all healthcare solutions, the biggest challenge, if you talk to all stakeholders in the healthcare space and you ask them what's the number one problem, they will tell you it's patient engagement. It's really difficult engaging patients to do something and that's why you have these opioid epidemic because it's very easy to take a pill, but that's not the solution that's actually a driver of the problem. So which ways can we make our solution more engaging and our thesis is that making it much more human-like in terms of you can talk to Phoenix, Phoenix changes what you are doing based on what you just told Phoenix. It provides actionable feedback, enables human-like feedback throughout the session in a highly engaging way. If you had your best PT cheerleader next to you at 4:00 AM in the morning, that’s our thesis that drives engagement. And so that was our thought process, why we decided to build it.

And then in terms of the internal machinations of making that possible, it's a quite complex system because you need to be able to measure the movement of the patient as the patient does the therapeutic exercises in a clinically great way. And for that you use different AI techniques which are not LLMs. You need to provide feedback to the patient in terms of understanding what you did wrong and how to improve. Those are different AI techniques and then when we use LLMs is basically voicing that feedback and then that's part of communicating to you. Then there's a part of you communicating to Phoenix and then you need to use then LLMs to transform the speech to text and then you need to use the different AI techniques to understand what you've just said and then you need to transform that into a different set of objects and actions from what you've just said. And then you need to change the program according to your clinical protocols because it's not just like, oh, Phoenix, I'm feeling a little bit tired and Phoenix replies, “So I just changed the session to be a little bit more lightweight today,” but it's not AI that is defining how the change is made. It follows clinical protocols that we have developed. So, it's a highly complex system where LLMs play an important role but it's not the most important role.

I'm curious, obviously it sounds like you've actually been able to do a ton with this product in terms of being able to adjust routines and do it off basic clinical guidelines to actually evaluate exercises as they're happening. What can't we do yet? What were some of the things ideally that Phoenix would be able to do but it's like either the models aren't good enough yet or there isn't good enough tooling yet to fully do some of the kind of end-to-end vision.

So, the thing that I've wrote, I wrote this on LinkedIn and on Twitter is what we're launching now is what I thought we will be delivering and launching five years from now. So right now there is nothing that I wanted to build on Phoenix that is not there yet. Now we need to polish some rough edges, but in terms of the technical possibilities of the solution, it really checks all the boxes of what we wanted to build. So it's really a marvelous time in that regard.

To be able to combine all of those different types of models together and have them all kind of working as well as it sounds like they do is incredibly exciting and I guess now you release it out into the wild and see how it's perceived, which I'm sure will be a very exciting few weeks for you guys. How does a rollout like this work? Are you doing a small number? Did you do a small number of people first? Are you kind of just doing a broad release at once, or are you trying to test out certain parts of it?

Yeah, so yeah, it's a staged release and a stage by solution. So now it's in the pocket of our population with Thrive, and Thrive is our physical pain solution, right? Then we are gradually expanding to the full population of Thrive. After we do that, then we launch in Bloom, which is our pelvic health solution and we launch in a pocket of the population there and then we expand. So it's a gradual, this is healthcare and it's not like, oh, you launched this, didn't quite work out, it's okay, let's revert to the previous version. People are expecting you again trusting you to recover and so that's why you need to be extra serious about it.

I'm curious how this scope of practice stuff works here. Does a PT normally change? Can anyone sort of change guidelines you give to a patient in terms of what they have to do next? Does a PT come in the loop at some point of this? Is there a human in the loop or is it full automation here?

So our AI care model is the combination of our AI care specialist Phoenix and the human care specialist. So it's always the combination of both elements, both stakeholders and the way this works, a practical example will be Phoenix always does the independent session with the patient, but then by the end of the session, Phoenix sends all the data to the cloud where then a member of our clinical team is going to access what we call the AI feed and there Phoenix acts as the copilot of the PT recommending changes to the program and recommending you just did your session perfectly yesterday, so Phoenix recommends me to send you this message congratulating you for being very good at that session and not having pain in that exercise that you had in the last two weeks. Now the important piece here is that it serves as the co-pilot where what this means is that every single action needs to be approved by the PT or rejected or edited, so nothing goes out without the PT approving that. And that's very important because, well, I will definitely not feel comfortable to have an AI system working, being live without guardrails.

So the care specialist works as the checks and balances of the AI system and also expands the AI system in terms of ability. And then of course Phoenix as the co-pilot really then scales and provides leverage to the PT. So it definitely knows how you felt in that exercise three weeks before and can use that data point to engage you much more saying, “Hey, you three weeks ago you had lots of pain, exercise, no pain.” Now that doesn't happen in traditional healthcare. That's a particular example where Phoenix as a co-pilot also is very, very powerful.

One thing I'm struck by is obviously some of the stuff you've built is bespoke to PT, like observing exercises, but obviously the idea of coaching and changing a clinical pathway or recommending changes to that based on what you've observed feels like it's relevant across the board. And so I'm wondering, I'm struck by you saying, Hey, I thought at first this would take us five years from now to build and now it's possible. Is this possible in other areas of healthcare today? How has that changed your thoughts on where we are today and also where you might go with the technology you have built?

So this is, you're completely right. Every single care model independently of the condition where it's 100% level intensive should be disrupted with this model. And of course you can figure out which other areas of care follow this model. And so we expect to expand our AI care model and really break this paradox of for it get one hour of care. You need one hour of this highly scarce human resource, which is the clinician and care specialist and want to break that paradox the same way we've changed the way we do things with physical pain, we change the way we do things with pelvic health, the way change we do things with sedentary behaviors, we want also to disrupt other areas of care.

How do you think about the boundaries of where you would want to go? Where is maybe not as interesting for you longterm? Obviously there's lots of other chronic conditions and areas employers are focused on. How do you think about those boundaries? I'm particularly interested in this question of do you go deeper down the stack of the MSK related stuff or do you go to other chronic conditions?

And the immediate answer is both. You do both. And we structured our organization and the org chart was structured in a way that allows us to execute on both independently where we going after pelvic health doesn't change our success with physical pain to be clear, just focus on physical pain and just expanding and getting more touch in there. In itself, we would build a massive company, it's 1.7 billion people all over the world. You don't need to have to access the 900 million women with pelvic health conditions to be a massive company. But we are mission driven and so if we feel that we have a solution to a problem that exists right now that is not being solved, we want to expand towards that. And so I think we've been so far successful building a company where you can look at as five startups in one where each one of those work independently and are able to pursue as horizontal expansion while we also are able to go deep down and vertically expand.

You mentioned a little bit about how the org structure sort of enables this. Can you tell us a little bit about just what does the org structure look like? I assume probably you've made changes as it's grown, but I'm curious why is your org structure unique in its design to be able to do those things?

I don't know if it's unique. So we built it from first principles but probably someone else got to the same conclusion. So we have the horizontal layers which are finance, HR, legal, engineering that support independent, we call them waves but are independent product solutions. And so one thing that we do well is let's assume that we want to expand into Condition X next month. What we do is we hire a GM, so a strong generalist past founder or that wants to be a founder; we hire a GM and we give that person $1 million to put a team and to execute and get to product market fit 12 months, that's your seed round. And of course then you are able to accelerate much faster because you have access to finance and legal and HR. You don't need to spend your time or focus thinking on that.

Then you have also an engineering team that you can recruit, but by the way you pay, it costs you money. So part of the $1 million is going to pay that allocation for the engineering team if you are just launching your startup. So we want to replicate that scarcity mindset and then if when after 12 months we discuss was there product market fit or not -- if there was you do your series A, which means that you get more money, you get to accelerate faster. If it doesn't, then we kill the solution. If you go to product market fit before the 12 months, you are able to raise your internal series A faster and then you expand. And one example Bloom, I can tell you Bloom was our second solution that we launched and that's the pelvic health solution. We are going to do as much revenue with Bloom in 2024 that we did with Sword all together in 2022. And so it started like that. Right now Bloom, if they were an independent company, they will be like a series C company, series C, series D company.

I mean I'm curious because obviously on your site you can see all these business lines you've launched and so it feels like this has been very successful for you. I'm sure there's been some of these that haven't worked out. Are there examples of things you guys have tried -- given that seed funding -- and it just didn't end up having product market fit?

Yeah, we are very open about it. One thing that we tried was get into the provider channel. We've been very successful with employers, very successful health plans. Look, we have a solution, let's go to providers. We've made a good effort. We were not successful independently of our effort and so it was clear that there was no product market fit and that we wouldn't be successful so we killed the play. And that in itself is success because we've learned that we should for the foreseeable future stay away from that. Yeah, but that's one of those that didn't work out.

As you talk about this broad product roadmap, I mean I think I heard you on another podcast talking about acquisitions and potentially that fitting in the future. How do you think about that? And obviously it seems like you guys are quite effective at building things zero to one, but where might acquisitions fit into this project?

So typically acquisitions are very successful when you have a good distribution channel and you don't have a product or an innovation muscle and you build products that are now starting to inject them in that distribution channel and you get an amazing ROI from that and that's a very successful strategy. We don't have that much of a good distribution channel. We don't have big partnerships with health plans. What we have is a good product and innovation engine. And so for us actually when we look at an acquisition is we get to get those distribution channels quite candidly. We've been looking at some things; we still haven't found one that could make lots of sense in that regard because usually we look acquisitions the other way around and that's why for us, in spite of some conversations that we are having, that we have in the past, we've never pulled the trigger on anything because in the end there was not a fit.

Just speaking about distribution really quickly, brokers sort of play obviously a huge role in this sort of ecosystem. Do you guys work with brokers or have you purposely strayed away from that? And if you do, how does it work?

So yes we do. They are important stakeholders in the benefit space and for you to be successful, you need to work with them. And it's pretty much a grassroots play because there's lots of local offices, you need to work with those local offices. Early on we put together a team just focused on what we call alliances. It's focused on brokers and consultants and really working with brokers and other consultants to really drive brand awareness and communicate our value proposition. And that's really key.

Going down the broker channel, how do you actually use that the most effectively? How do you actually do a good job with your brokers?

I think we do an okay job. I think we do an okay job with our brokers. I won't say we are stellar at it. I think we do an okay job. I think with brokers and consultants, I think it's the same thing with clients, but I think it actually it's even more important to them to bring data, clear data so that you are not another company saying, “Hey, we are super cool. We have a net promoter score of 200 and we save you, reduce you all of your healthcare expenses.” It's important to have clinical validation, it's important to have proper data so that they take you seriously. And so I think that's the most important thing. Again for me, all of that stuff, clients, fundraising, brokers, consultants, partners, you need to do the work. You need to do the work of building something that is valuable and then having proof points of that value. So especially in such a noisy market, I think that's key.

What other pockets of the international healthcare opportunity are interesting to you? Are you guys live in Portugal?

We are live in Portugal. Not in the national health system, to be clear why? I'm not sure, but we are not. We have more than 10% of the Fortune 500 listed clients, more than 10% of the Fortune 50, biggest payer in the US, one if not the fastest growing health tech company in the US in 2024, in Portugal. The nation itself is still asking us to do a pilot to prove, to showcase our value. And so it's clearly a lack of understanding and honestly we don't care because we are going so fast in the US that we couldn’t care less. But it really clearly showcases how challenging it is to build a healthcare business, a novel innovative healthcare business in Europe. We have a global presence. We are in Canada, we are in UK, in Portugal with insurance companies. We are in Australia and we are expanding globally, but honestly we are expanding globally because our US clients are asking us to expand globally. So we are supporting our US clients in Canada. We are supporting our US clients in the UK.

We have other areas and countries where we are going to expand into 2024, but it's really a shallow penetration supporting more US clients, global US organizations than really a deep penetration. This being said, we have one of the waves is focused on the international segment. We are very focused on that, but it's really a long-term play and I don't expect for that to meaningfully represent a good portion of our revenue either in 2024, 2025, or 2026 – it’s really thinking toward as this generational company that is going to be here 50 years from now and it's going to take time. So might as well start now.

Do you notice any interesting different behaviors from patients in the US versus patients in other countries or does everyone just kind of use it the same?

The reason why we are so focused on international expansion is because it's the same thing. Treating a member in India, treating a member in Germany, in UK or in US is exactly the same thing, exactly the same pain, exactly the same struggles and exactly the same solution to address those struggles. So what changes is actually the most important part is how you get to patients, how you get to people, and that changes a lot.

Totally. Well V, we always like to end our interviews with a quickfire round where we get your take on a few common questions we ask guests. And so maybe to kick things off, we’d be curious for one thing that's over hyped and one thing that's under hyped in health tech right now?

I have the same answer to both things. AI is both over hype and hyped. I think in the short term it’s over hyped, I think in the long term it’s under hyped.

So second question is, and you might have an interesting perspective of this, you're kind of international and you're looking into the US healthcare market with an outsider's lens. Is there a piece of regulation or some rule that when you started selling in the US market that you were like, this makes no sense, we should change this?

Many one that still to this day, I'm like why? It's the fact that a PT in California cannot treat a PT in the next state. And that for me, it's like it's the same country, it's the same PT, the same American citizens, but because one is located in a different state, then you are not able to do your work. That's very weird. And that doesn't happen in Europe. You have nurses in Portugal that are in UK doing the work without any kind of certification, but for some reason in the US you have this thing. And so we have a machine internally to enable our clinical team to provide care in 50 states. It's another overhead for something that I don't believe makes sense, but it exists. So we comply with it.

I was listening to a podcast with Vinod Khosla and he said that you flew out basically to have dinner with him and it was this very impactful dinner and I left very intrigued. It sounded like this big, big business conversation. I was just curious, anything you can share about some of these kinds of conversations you've had? Obviously Vinod is such an interesting person in the broader health tech ecosystem, but this dinner, I guess the way he told it, you flew out and kind of flew back right after this dinner and it was like this big thing. So I would love to hear a bit more.

Yeah, so basically Vinod texts me, “Hey, do you want to have dinner next week?” And I'm like, sure. And I'm in Portugal, I need to catch a 20 hour flight to get to San Francisco to have dinner and then get back in another 20 hour flight back to Portugal. And I do it every time quite happily because it definitely pays off because, again, Vinod is one of my heroes and the inspiration that you get from those conversations, the ambition, and I'm a very ambitious audacious guy, but he basically is in a different dimension and the way he pushes me to think about things in a contrarian way, but even more ambitious than I thought previously, clearly pays off that 20 hour flight. And by the way, it's not that bad to be 20 hours to be meeting anyone because you can do lots of work, especially when you have two little kids at home. So it's like there are other side benefits on catching a fight to just have dinner with Vinod in San Francisco.

As a follow up to my last question, I'm wondering, obviously there's so many little things over time that have made your product so compelling, but as you look back over the last nine years, is there one to two big decisions you can point to where you're like, wow, besides writing the cold email to Khosla, which seems like it was a game changer, are there other forks in the road where you look back and you're like, going in that direction versus the other one was really game changing for the company?

There was one transformational moment for us, which was when we started, our first vision was to develop this technology for traditional PT teams in a PT clinic to use it. And so basically the patient will go to the PT clinic, they will have the first session there, then go home without technology and do the sessions independently at home and they will be remotely supervised by the PT. And we thought this is quite a clean model, clearly extends the footprint, blah, blah, blah, blah, blah, blah. And then we launched a pilot with some clinics, and PTs were complaining all the time that the system, and this was like 2018, that the system was too buggy, were always lots of problems that patients couldn't use it, they couldn't take it. And we were like, well, it's not like we are naive and we think our system is perfect, so one day we decided, let us try to have a pilot with patients directly at home and with our clinical team.

So we hired some PTs to do the work. And funny enough, the system worked beautifully in the patient's home. There were no bugs whatsoever. And then what we've realized is that there was this Clayton Christensen innovator’s dilemma going on where PTs were like, oh, this is going to replace me, so I don't want for this to be successful. And they were complaining about the system all the time, but we were very afraid to go directly to the patient's home because we thought, well if this is buggy and it's not working in the clinic, which is a safe environment and a controlled environment, how is it going to work independently in the patient's home without any clinician there? And we were very afraid, and I know that I had a very big discussion with my chief clinical officer where he was really against it and he still owes me a dinner to this day that he lost that bet, because then we bet, look, I bet that this is going home and this is going to be successful.

And the reason why it was successful right from the gate is because it's not like the system was perfect, but even with all the imperfections and the rough, rough edges, it was much, much better than going to a clinic, asking your wife to get you there, find a parking spot, and then doing everything back home and then waiting in the waiting room 20 minutes, even if the system wasn't perfect, it was 10 x better than the alternative. And so right from that moment we understood, oh, okay, so we are not going to be a company that provides technology to existing teams. We are going to be a provider of care using AI. And so that was actually a big inflection point in our history.

I feel like that's the perfect note to end on. I guess I want to leave the last word to you, since there's going to be tons of threads people want to pull on. You guys obviously are announcing exciting products. Where can folks go to learn more about you? The floor is yours, anywhere you want to point people.

We have our website, we have our social pages, I have my Twitter and LinkedIn. And of course people can also send me emails directly. My email is quite straightforward, it's just v@, and so please feel free to reach out.