Medsider: Learn from Medtech and Healthtech Founders and CEOs - Connecting the Dots Between Medtech and AI: Interview with Sonio CEO Cecile Brosset

Episode Date: November 27, 2023

In this episode of Medsider Radio, we sat down with Cecile Brosset Dubois, CEO of Sonio, a company dedicated to revolutionizing obstetric ultrasound through AI technology.  Cecile began her career in 2005 at Capgemini, later advancing to Bain & Company. Her strategic insights as Head of Strategy at FSI - Fonds Stratégique d'Investissement - and her pivotal roles at Bpifrance set the stage for Cecile's entrepreneurial journey. At Kamet Ventures, she fully embraced the startup world before eventually co-founding Sonio. In this interview, we touch on her strategic approach to product development, the importance of an in-house regulatory team, her drive for meaningful work, and effective tactics for fundraising. Before we dive into the discussion, I wanted to mention a few things:First, if you’re into learning from medical device and health technology founders and CEOs, and want to know when new interviews are live, head over to Medsider.com and sign up for our free newsletter.Second, if you want to peek behind the curtain of the world's most successful startups, you should consider a Medsider premium membership. You’ll learn the strategies and tactics that founders and CEOs use to build and grow companies like Silk Road Medical, AliveCor, Shockwave Medical, and hundreds more!We recently introduced some fantastic additions exclusively for Medsider premium members, including playbooks, which are curated collections of our top Medsider interviews on key topics like capital fundraising and risk mitigation, and a curated investor database to help you discover your next medical device or health technology investor!In addition to the entire back catalog of Medsider interviews over the past decade, premium members also get a copy of every volume of Medsider Mentors at no additional cost, including the recently launched Medsider Mentors Volume IV. If you’re interested, go to medsider.com/subscribe to learn more.Lastly, if you'd rather read than listen, here's a link to the full interview with Cecile Brosset.

That would be trust yourself. Believe in yourself, believe that you can do it. Don't let other people distract you from your objective and just do it. Just to kill them. Welcome to MedSider, where you can learn from the brightest founders and CEOs in medical devices and health technology. Join tens of thousands of ambitious doers as we unpack the insights, tactics and secrets behind the most successful life science startups in the world. Now here's your host, Sky. Scott Nelson.

Hey everyone, it's Scott. In this episode of Medsider, I sat down with Cecil Brossette de Boy, who kicked off her professional journey in 2005 as a junior consultant at Cap Gemini before joining Bain and company in 2007. After nearly five years at Bain, Cecil transitioned to FSI, serving as the head of strategy until 2013. Following this, she played key roles at BPI France, initially as director of innovation development and later as director of BPI France's Le Hub.

After becoming an entrepreneur in residence at Comet Ventures, Cecil co-founded so and leads the company's efforts on enhancing obstetric ultrasound with AI. Here are few of the key things that we discussed in this conversation. First, the guiding principle both in product development and team building is the same. Progress is iterative. Aiming for perfection while a worthy goal should not paralyze the journey towards improvement. Be tenacious and don't give up. Second, having an in-house regulatory team that understands both the product and the ever-changing landscape is a must,

especially if you're utilizing new AI-based technologies. Third, whether you're commercializing or fundraising, understand your audience and keep an open dialogue with experienced entrepreneurs. Focus on building relationships and leveraging them for future growth. Before we jump into this episode, I wanted to let you know that the latest edition of MedSider mentors is now live. Volume 4 summarizes the key learnings from the most popular Medsider interviews over the last several months with folks like Rob Ball, CEO of Shoulder Innovations, Kate Rumrell,

CEO of Abolative Solutions, Dr. Christian Ramdo, C. CEO of Tempe Health and other leaders of some of the hottest startups in the space. Look, it's tough to listen or read every Medsider interview that comes out, even the best ones. But there are so many valuable lessons you can glean from the founders and CEOs that join a program. So that's why we decided to create Medsider mentors. It's the easiest way for you to learn from the world's best medical device and health technology entrepreneurs in one central place.

If you're interested in learning more, head over to MedsiderRadio.com forward slash mentors. Premium members get free access to all past and future volumes. And if you're not a premium member yet, you should definitely consider signing up. In addition to every volume of Medsider mentors, you'll get full access to the entire library of interviews dating back to 2010. You'll also be able to see all of our playbooks, which are handpicked collections of the most insightful interviews with the brightest founders and CEOs. Whether you're looking to master capital fundraising, navigate early stage development,

tackle regulatory challenges, understand reimbursement, or position your venture for a meaningful exit, MedSider Playbooks have you covered. And last, considering that fundraising can be one of the best, most daunting task for any startup. We created a meticulous database of investors right at your fingertips. Explore a wealth of VC funds, private equity firms, angel groups, and more, all eager to invest in medical device and health technology startups. Access to this database is a premium member exclusive, so don't miss out. Learn more about MedSider Mentors and our premium memberships by visiting Medsiderradio.com forward slash mentors. Again, that's Medsiderradio.com

forward slash mentors. All right, without further ado, let's jump right into the interview. All right, Cecil, welcome to Medside Radio. Thanks, Scott. Happy to be here. Yeah, looking forward to the conversation. I know you're kind of over in the States from Paris, where I think the company is based. So good timing for this conversation.

So on that note, I recorded a very brief bio of your background kind of at the outset of this interview. But let's start there. If you can kind of give us an elevator pitch for your professional background, kind of leading up to starting, Sonya, that'd be great. Sure. Happy to do so. So my background is actually a business background.

I've started with, you know, almost five years of consulting, working for a band and company, you know, kind of keeping some learning some stuff, you know, before I going into like the real world. So that was super helpful. And then I really wanted to give some like meaning to my work. And I come from a family of doctors. I've always wanted to have an impact. And so that was kind of my criteria for my next job after,

after strategy consulting. And I joined the French public investment bank where my job was to actually, you know, finance startups or actually make sure that we could find ways to help them grow. And so that's when I got to, you know, meet with tons of entrepreneurs and understand, you know, what were there, the issues or the thing they were struggling with and how we could help them. And around that time, actually I saw, you know, those. amazing, you know, health tech startups emerge. And I always saw that, you know, startups could only do like biotech companies.

Like so, and then, then R&D would take quite a lot of time before being able to have an impact on like the life of patients. And when I saw that, you know, with AI, you could actually, with technology, have an impact on health quite, quite rapidly. I love that. And that's when I thought, you know, I really need to get my hands dirty now. And, you know, stop watching those entrepreneurs, but really. bring some value with, you know, myself.

And yeah, and that's basically how I ended up, you know, meeting with some great folks and building Sonia. Got it. That's very helpful. And I think I'm looking at your LinkedIn profile now. So you spend about four and a half years, it looks like at Bain and then another almost six years at BPI, France. Yeah.

Right. And so we're recording this in, you know, let's call it early, Lake Q3, I should say, of 2023. And you've been at Sonio now for almost four years. about right? Absolutely. Okay, cool, cool. I just want to kind of set the stage for everyone listening. So it looks like you started Sonio in late, kind of late 2019. Give us a sense for kind of what it is you're working on, right? And then also maybe the Ors the story of how you, you know, how this technology sort of came to be, how you identified it and kind of where it's at now.

Sure. So, so yeah, Sonio in a nutshell, it's an AI software for obstetric ultrasound. So basically what we do is we try to empower OBJNs, MFN, sonographers that do, you know, those scans for pregnant women. The thing is, and many people don't know about that, but not all pregnancies actually go well. That's why we do those screening examinations and 50% of malformations that might affect, you know, the fetus actually not detected during those ultrasound examinations. So 3% of babies, you know, actually are born mental defect and you still have 100,000 dying, you know, every year from, you know, causes that, that could be, you know, preventable if detected during

pregnancy. So those are kind of the numbers that, you know, prevents us to from like, sleeping at night. And we really thought that, you know, bringing the software that could be kind of the daily companion of those practitioners might help them, you know, again, be alerted in case something is wrong. And also just, you know, improving the daily routine, the level of digitalization, you know, is not, it's not super high for doctors. It's kind of shocking how obsolete the software they work with are. So we had, you know, kind of a true mission here, you know, that we needed to address. And basically, AI really helps democratizing kind of quality standards for everyone in this field. That's what we're like running after.

So how did I end up in this field? I said I wanted to find a project that was kind of a digital health project. And I just met with a lot of people. And I finally met with these amazing doctors and mathematicians at the end of 2019. And there were references, you know. So NECA is the main hospital in France and one of the most important ones in Europe where you have a level of fetal medicine that's incredible. So basically a lot of the babies or the prime women have had helped

issue that are referred to the center. And I just met with the Professor Eviel. He's the head of the department of the Joanne and MFM there. You know, so a brilliant person. And I met with, you know, the people he was doing research with for like over the last, you know, five years, especially Remy Besant, who became my co-founder, chief scientific officer. that was just finishing his PhD, working on, you know, like recognizing, helping to diagnose these,

you know, abnormalities and putting together diagnosis. So, you know, just, again, it's always about, you know, meeting the right people. And then it was my job to kind of understand whether or not there was enough space for, like, building a business and how to build the right product to meet the needs. That's why we did. Got it. And so I'm looking at your site now. It's sonio.a.i. So S-O-N-I-O-A-I-I-I-I-I. And so you're, for those listening that don't make it to the full summary of this interview on MedSiter, you can definitely check it out. That's sonio.

AI. So you're providing just the software alone, correct? Absolutely. Yeah, yeah. We are not building any hardware, but we're compatible with any ultrasound system. I think this is part of, it's one of our main values. We are vandal neutral. And we're bringing this like workforce solution.

So basically we're replacing. the current reporting software that, you know, practitioners are using connecting to their AMR and to their billing solutions. And at the same time, we are connecting real time with those ultracea systems so that we can collect the images and make sure that, you detect,

the practitioners take all the right images. They don't forget any. It's like fetal, you know, imaging is quite, you know, tedious. You need to take like 20 to 60 images. You know, you have 20 to 30 minutes to do that. And you need to make sure to acquire the images correctly, you need to interpret them correctly. And again, then you have a reviewer like a reader

that needs to make sure that, you know, they look right and there's no, there's no issue. So it's very operator-dependent. And so having this ability to bring some efficiency throughout the process, because they all have to document, you know, their examination. So building the report, you know, automatic or semi-automatically, thanks to AI, but also having life feedback if something is wrong with the image or an image is forgotten, has. really lot of value for them, you know, that they do work. Got it, got it. Yeah, my wife and I have five kids.

So I've sat in on a number of these types of scans for sure. So definitely not an expert, but I've seen plenty. So if I understand this right and pretend on maybe like a, you know, freshman in high school, I just want to summarize, I think what I understand. So you're providing the software that aids in probably an efficient collection, right, of scans, right, during the actual, you know, ultrasound. sound kind of lab, if you will. Is your AI also, your software also providing assistance to the actual physician in

terms of diagnosing any abnormalities, any issues as well? Yeah, absolutely. So the idea is really to bring like an end-to-end solution and really answer all the pain points of the physicians and the sonography. So the most important thing is really to help this screening process, making sure that like everything is okay. if there is something unusual that's detected on an image, then what happens is that usually you will be 211 or will refer you to an MFM

or to like a fetal medicine center that we'll need to make sure that there are no other, like that there is an abnormality and make sure that there are no other abnormalities like that is not like something like a rare disease or something that's logical. So you need to really interpret the malformation or the sign. itself know whether or not you can do something about it because some malformations can actually be taken care of through surgery either during the pregnancy or at birth, but you need to really, you know, take care of it right away. But so if you have a collection of science, you need to, you know, again, know, know what type

of disease it might be. So you analyze the phenotype. And I have kind of another AI. So the first AI I described to be able to, you know, automate the report and make sure we have all the images and they're okay. It's a deep learning algorithm. So basically, you know, you learn from a lot of images. Now for helping with diagnosis, we built another type of AI, like you can't symbolic

AI. It's an expert system if you want, the probabilistic, where we try and mimic the clinical reasoning of the practitioner. We help him in a probabilistic way. We help him look for associated science and then we tell him whether or not it's a genetic disease or infectious disease or if it's just an isolated anomaly. And so it changes the way you will take care of the patient, obviously. Got it. Got it. Okay. So almost two different aspects of the software, it sounds like. And you mentioned earlier, you met these physicians,

your co-founders, right? Your physicians and PhDs in late 2019, we're recording this, again, in Q3 of 23. So almost four years later now, where's the company at in terms of life cycle from a kind of a development, you know, Clint Ray commercialization, etc. Yeah, we grew quite fast. Now we are a 30 people team. And, you know, beyond the first co-founders that we talked about already, after like a year, I actually added three amazing, you know, co-founders to the team, late founders. So David, who was the CPU and is now the C-O because we really needed some great product skills

to make sure, you know, we can bring like a software, medical device, you know, together. Deepak Prakash, who's like the CTO, who's kind of, you know, really from the digital, you know, world and can build state of the art technology, you know, architecture, real time, you know, software, which was really important for us. And Dagmanuba, who's the, she's a three third late founders to join. She has 20 plus years of experience in the field in over, I know, reporting for obstetric ultrasound. She built viewpoint, it's kind of leader in the market, and then sold it to G. So, you know, we first built this kind of core team, kind of, you know, year after founding the company. And then, you know, we financed the company three times already and made

sure we could reach out different goals. So first goal was to build the product. And now we are the stage where we have finally kind of the full platform that I'm, that I've described. So the workforce solution, the deep learning AI to really automate everything and do the quality assurance and then the diagnostic tool. So we really. we have this full platform ready to be commercialized in the US. We got FDQIED a few weeks ago, which was kind of a great milestone that we achieved. And we have a CEMARC for the diagnostic tool in Europe.

And now we're going to actually go and see Mark the deep learning AI in the coming months. And so, yeah, in a nutshell, we are kind of a product company, tech company that is going to, you know, changing into a commercial company. And that's where we are after, after three and a half years. Okay, cool. Yeah. So made a ton of great progress in a reasonably short amount of time, especially considering your technology, there's not a low barrier here, right? I mean, it's pretty sophisticated, pretty sophisticated software.

So congrats on that. That is, that's really, I always love to see sort of life science, med tech companies, like moving fast, right? Because you rarely see it. You know, I often, oftentimes run into a startup that's like 15 years old. and it's like, whoa, what are you guys been doing for this long? So really cool to see how fast the team is moving. So with that said, let's spend maybe the next 20 or 30 minutes kind of covering sort of the key functions, right, that any startup has to work through, right,

to get to the point that you're at. And I know you mentioned the three financings, which I'd love to get your take on here a little bit later in the conversation. But let's first start with kind of just product development in general. I'm sure the first versions of Sonio, your AIA platform, looked a lot different than what you're, sort of commercializing now. And AI is no small feat, right, from a technology perspective. So give us a sense, you know, from your perspective, especially considering you've worked with and

seen so many different startups, you know, both at Bain and then at BPI France, what's your advice for how other founders and entrepreneurs can be, can stay, can be really capital efficient, but yet still move fast from a, from a development standpoint. Any kind of key lessons learned, you know, based on your, your professional journey so far? Yeah, good question. So just, to answer the first part of your question. So yeah, of course we didn't build the full platform in like one go. We started by some modules in a way that were also like tested all the way through. So basically we started with this really niche product, this diagnostic tool. That's a kind of a technical assistant support tool. So and we basically we got, we tested everything

with this first module. So we tested, you know, building the product itself. So combining AI. I was, you know, just like, you know, software development and then making sure we could like find people to use it. And we went through CEMOC with this first module. And it was great because we learned so many things by like doing so. And then we started developing some modules of the workforce solution. So we, we developed the checklist and we developed the like sharing some images. So those are kind of features now of the full platform. But this really helped us, you know, go fast. to market and really get like iterate with like feedback of the customers and make sure we also identified who we should target in the market. What were the expectations? You know, and so it's really

informed our roadmap a lot and it also informed our decisions to like recruit or, you know, how to structure the team and how should the different, you know, teams inside the team work. Like, for example, what decided to create kind of a BU that would be products that would be, you know, that would go from data science to product and operations. So data science, clinical regulatory product under the same umbrella so that we could really anticipate what would be expected from, you know,

like the regulatory, from a regulatory standpoint early enough in the process. That would be product and regulatory that would inform the data science roadmap and the tech roadmap and the clinical validation. And so these are all great learnings that we've done by iterating again on small modules before actually building the true product that we are now starting to commercialize.

And I think, you know, I had heard that so many times, you know, when I went at BPI that you have to iterate and, you know, doing it, you know, without doing it perfect is what really, you know, is right. And I think, again, I could experience it. And the other thing that I think is that it was really important to me and that I think this is one of the amazing thing that happened is that we did the right choices in building the team. We'll be identifying the right people that would be key contributors so that we could have an amazing software and an amazing medical device and amazing, you know, so amazing clinical, from a clinical perspective, from an AI perspective, from a software perspective.

and like adding those different skills to build, you know, a best in class, you know, product. And that that team structure that you mentioned earlier, did you build it around, it sounds like if I understand it correctly, you built it around that process, right, from product to reg that would inform. I think you mentioned kind of data science. My understanding, it sounds like you almost set up a cluster of people in that same, under that same umbrella then. Is that right?

Yeah, absolutely. Absolutely. At some point, we realized that product couldn't. just be kind of, you know, the team that would manage, you know, the tech roadmap, but that really the constraints were too strong on the AI modules. And so that we really needed to anticipate quite a lot. So the great thing about our platform is that some modules are not regulated and the workflow solution is not regulated. So we could really, you know, develop this continuously in a trade with our users. While the AI modules were.

regulated. And so we really had to take into account all the constraints early enough in the process so that we would do the right things. And yeah, at some point, that's why the CPO actually became C-O, because he really had to be managing the whole process early enough so that we wouldn't develop the algorithms the wrong way or in a way that wouldn't again suit the regulatory constraints. And same thing. So that was kind of a mix of listening, of course, the users and their needs, but also again, managing regulatory. And regulatory, there's always room for interpretation, you know, so that's when you need to be tactical a bit.

So you need to also define the product in a way that you can answer your, you know, the user's needs. But at the same time, maybe not be, not go too far because then you need to do, I don't know, I'm just going to say, you know, whatever, the prospect of study, because your claim is just too high. You need to, we didn't do prospective study. We thought, okay, so we don't want to fall in that category. So what can we do that will bring enough value to our users,

but that would still allow us to put the product on the market, soon in the next like a year or a year and a half. And so we made those choices to not, you know, not ask for too much too fast and then just be able to buy little. Yeah, yeah. I want to circle back around to kind of some additional kind of insights around that topic, right? Sounds like you've got some. great things to share, but certainly back around to kind of that team structure. Did you kind of

understand that how you wanted to build out that team out of the gate? Or did you sort of learn early on, oh, wow, we need to kind of, we need to make sure that all of those folks are not silo, that they're on the same sort of team or sort of under the same umbrella from the gigo? No, it was not something I knew from the start. At first, I just had, I hired the right people for for their scope. I really picked them for their expertise. So for example, I knew from my past experience that like product management was absolutely crucial. And so and again, and I didn't want to have also the, the CEO come from the healthcare world because for me, at least in France, the, the market was not mature enough. So we would have, so I would find the best guys in this industry. So I'd rather,

I really wanted to go for like a guy that I would know again how to be all the amazing software. So I think this is what I tried to apply, like pick the right talent in his or her specific field. And then it was my job to make sure that they would adapt and understand what was specific about our business and then learn from each other. So that's what I did first. You know, I kind of, again, added them all. And then, of course, then you have, then leaders, you know, emerge naturally. Like a funny story about David Demoyal to my CEO now. And it was kind of my, really my partner in drive.

you know, he's the one that I managed the team with. I first, I didn't know, I was not 100% sure when we met that he would be the right person. And but he said, you know, I believe in the, in the, in your idea, in your vision, in the team, I really want to do that. He didn't have like a healthcare background. He was an amazing CPU at different startups. So try Tio first and then other like French startups, legal start, you know, tiny clues. And but he said, okay, I believe in the project.

I believe in you. I'm ready to come as product manager to start with. And again, decrease his salary, his expectations by a lot. And I had, okay, I'm sure I will convince you in six months that I can be CPO.

Like, okay, let's do that. And so we tried it out. And, and, you know, it was amazing.

And then he was promoted CPO, you know, six months. And then again, a few, a few, a month, years later,

then again, we realized that C.O. would be actually the best position for him in our situation. So I think, yeah, it's just like trying to to work on people's skills and see what are the natural kind of dynamics that you create in a group. I really, that's what I love, I guess. Yeah. I love that concept. I kind of like to refer to it as like professional dating, right? Where you, I mean, you kind of did it with full time hires. You don't have to have like full time hires. You can do this even with consultants, right? Before maybe even you bring them on full time.

is like you there's this period right maybe it's three months maybe it's six months something like that where you get to see not only how you work together but also to your point you could see them maybe merge or breakthrough and and they can maybe either quickly become someone that's like oh that's going to be my chief my coo my cpo whatever right or or vice versa right it's just not it's not a good fit it's not working out maybe they don't have the ambition whatever but i i like that kind of that uh sort of that testing period if you will and it sounds like that that that's worked out for you. Sezio, it seems like you have kind of a knack for not only identifying really top tier talent, but also kind of winning them over and getting them onto your team. What's worked

well for you in that category? Because I mean, you're working with people in the AI space that are hard to come by. I mean, AI is like literally arguably, probably inarguably the hottest kind of sort of category right now, especially in the startup world. So you've been able to kind of move exceptionally fast, build out what seems like a really, really solid team, like any, what's worked well, I guess, for you in terms of attracting that, those kind of folks. Hey there, it's Scott, and thanks for listening in so far. The rest of this conversation is only available via our private podcast for MedSider Premium Members. If you're not a premium member yet, you should definitely consider signing up. You'll get full access to the entire library of

interviews dating back to 2010. This includes conversations with experts like Renee Ryan, CEO of Cala Health, Nadine Miarid CEO of CVRX, and so many others. As a premium member, you'll get to join live interviews with these incredible medical device and health technology entrepreneurs. In addition, you'll get a copy of every volume of MedSider mentors at no additional cost. To learn more, head over to MedsiderRadio.com forward slash premium. Again, that's Medsiderradio.com forward slash premium.