How to Sell to a Conservative Industry

SaaS is Eating Software

Interview with Deep 6 AI Founder & CEO Wout Brusselaers.

Show Notes

Among other things, the COVID-19 pandemic has taught us the importance of clinical trials. The more streamlined and efficient they are, the more lives can be saved by new treatments and devices. Fortunately, tech startup Deep 6 AI is using artificial intelligence to accelerate patient recruitment for clinical trials and make the entire process more efficient.

Deep 6 AI founder and CEO Wout Brusselaers joined startup coach Roland Siebelink on this week’s episode of the Midstage Startup Momentum Podcast to talk about how Deep 6 impacts the world and what he’s learned during his startup’s journey thus far.

  • How Deep 6 went from an AI proposition to being integrated into healthcare workflows.
  • What it’s like selling new software to an industry that’s reluctant to accept change.
  • The best advice Wout has received on making B2B sales.
  • How Deep 6 is trying to attract talent for an increasingly remote workforce.
  • The traits that Deep 6 values in potential team members.
  • The framework that Deep 6 has created for the things that matter the most.


Roland Siebelink: Hello and welcome to the Midstage Startup Momentum Podcast. My name is Roland Siebelink and I'm an ally and a coach and a helper for all kinds of fast-growing startups and founders around the world, one of whom is in my studio today. His name is Wout Brusselaers and he is the CEO and founder of Deep 6 AI based out of LA. Hello, Wout. How are you doing today?

Wout Brusselaers: Hi, Roland. Thanks for having me. I'm doing great. How about you?

Roland Siebelink: Excellent. Very well. We were already doing the pre-conversation in our joint common native language of Dutch/Flemish, which was very refreshing, but we did decide to switch back to English so that the rest of the world could understand us as well. Wout, tell us all about the company that you founded and that you are representing here, Deep 6 AI. What do you do in the world? Who do you do it for? And what difference are you making?

Wout Brusselaers: Sure. Deep 6 AI builds artificial intelligence software to help match more patients to clinical trials faster and with less effort. Why does that matter or who cares? As you may not know - or some people will know - any new drug, any new medical device, or any new procedure in healthcare has to go through clinical trials to make sure that it works and that it is safe. This is the application of the scientific method to healthcare. You want to test things before you let them go out into public. This is a great gatekeeper in principle and it's a wonderful idea. But in recent years, it has become a serious bottleneck to innovation and healthcare. And one of the reasons is that over the past few decades, we are more and more moving into what is called personalized or precision medicine, where new treatments are increasingly optimized for select subgroups of patients based on specific elements of their genotype or phenotype. Do they have a specific mutation? Do they have a specific disease? Do they have a specific lifestyle? But the tools that we have to precision match those select patients to clinical trials, they have not kept up. Instead, it is still the same way that we used to find patients for a clinical trial for aspirin. Aspirin, as long as you have a headache or you have some kind of a vague ache, you can take an aspirin, you'll probably feel better. That is very different from specific immunology drugs that target very specific molecule interactions with genetic conditions, et cetera. The same tools and methods that we could use to find any patient, pull them off the street, and get them into a clinical trial are no longer viable today.

Roland Siebelink: Okay. This is within the trend of going toward more specialized medicine; medicine that only applies to specific subgroups, as I understand it. What does AI do to help you overcome that problem?

Wout Brusselaers: It's a great question. As you may know, over the last couple of decades in the US and worldwide, we have actually pumped billions of dollars into the datafication of healthcare. We try to digitize most of the patient's medical records in a way that is shareable and consumable. What you would hope and expect after such a tremendously expensive and time-consuming effort is that you have massive amounts of big data. We can have all of this patient data and it talks to each other, you can compare one data of a patient to another patient, and it yields all of these new insights. Unfortunately, that is not the way of the world today. Instead, what you really have at the tail end of this tremendous effort is massive amounts of little data, locked and fragmented in individual patient records, and they scarcely talk to each other. And one of the key reasons - apart from potentially politics and business reasons - one of the technological reasons is that most of the salient data that gets into a patient record is what we call unstructured. It is data that is not labeled. You don't really know what's in there. And this is typical of physician notes. If you go and you see your physician and he asks "Hey Roland, why are you here? What is wrong with you?" You start talking about your symptoms, about your aches, about your lifestyle, about what you're doing. And the physician is trying to maintain at least a modicum of face-to-face interaction with you. He's talking to you and then he sometimes starts typing stuff on the keyboard. What he's really doing is filling out a long list of conversational language, where he's typing up everything you say, typically in one long text block. That text block is hard for traditional software to understand. Software likes it when you say: "Well, this is a symptom, and this is the name, and Roland is his first name and Siebelink is his last name, et cetera." If I just start typing away, software doesn't really know what is the name here, what is the symptom, what is the disease, what am I supposed to do? That means that much of the data you're putting on that record is not easily retrievable or actionable. What is our solution? We bring AI and NLP to that unstructured data. We're trying to actually make sense of that data by doing both natural language programming to say "What of all of this blob of text here is a name, and what of this is a symptom, and what is a treatment, and what is an outcome, and what is a measurement, and what is potentially a genetic mutation?" we recognize all of that in the text. We do the labeling on behalf of the physician or behalf of the health system. And once we do that, we can do other types of machine learning and AI to start training our models to recognize these things. To make sense of that data, you also have to know how relevant is the data. You can mention a clinical concept, but is that concept active? Does it pertain to the patient or somebody else? Is it of this day or is it historical? You use more and more of that modeling to contextualize it as well and to then drive certain decisions based on that. And that's what we do. We mine all of that data. We represent patients as the sum of all of that structured or unstructured data. We make it very easy to navigate. And then we build rules and decisions that allow you to easily identify patients for clinical trials and help you recruit them through the process.

Roland Siebelink: Okay. That sounds very compelling from a technological point of view. The question, as always with AI, is how do you get all the data? And that's often closely tied to who do you serve with this proposition?

Wout Brusselaers: Yeah, that's another good question. As you know, our name is Deep 6 AI. When we started a couple of years ago, we thought it was important to put that on the record. Because we really came with a deep background in AI, almost more than in healthcare at the time. But what we found is it's equally important to really focus on the UI. But it's really important, like you said, to make sure that when you have AI, that you solve a problem with it, and that you use the right data at the right interaction to empower users to make better decisions faster and drive results. The AI is the core that we build on and we train that against all of the clinical data that we pull from hospitals. Hospitals are often the executive branch for a lot of these pharma companies, medical device companies, who want to bring their new drugs or their new treatments to market. They rely on academic medical centers and other health centers to execute all of their studies. You recruit a patient, test these new devices, record how the patients are interacting with it, and then report that back to pharma. Because they're already in the business of taking care of patients. Doing that with novel therapeutics and seeing how well they work compared to others, it's a logical extension of what they do.

Roland Siebelink: You're basically jumping on a way that the value chain has already been configured in the past, knowing full well that pharma companies actually often don't execute that research themselves but outsource it to hospitals - academic hospitals, in particular. Then you're using these as the sources of your data.

Wout Brusselaers: Yeah, absolutely. And I love that you use the term value chain. I think it probably says something about both of us that we use these terms in conversation. That is really what we do. And part of why I brought up the concept of UI compared to just AI, is that particular to clinical trials, that whole value chain is fragmented across multiple stakeholders. You have the pharma companies, and it's their drugs and their devices that they want to bring to the market. They start a clinical trial and then they have to select sites - hospitals to work with to recruit patients, gather the data, report it back to them. They then submit that to the FDA, the FDA approves or does not approve the new drug, and then they can start selling it. That whole process, you have to interact with multiple stakeholders within and between organizations. And what we really do is we try to connect all of those stakeholders on an AI-driven, real-time, real-world data platform, so everybody can work together to make better decisions faster.

Roland Siebelink: Maybe this is a side story, but some people tell me there has been a little bit of a pandemic going on in the last two years around the world. And I wondered is that something that has completely exploded your business or because it was something that is so universally applicable that it was actually more of a break on your business?

Wout Brusselaers: It's a great question. And I feel like the answer in our case is mixed. We are selling to health systems and we're now starting to sell into life sciences companies. The executive arm of the clinical trials and the overarching rights owner arm of clinical trials. And in 2020, when the pandemic really started, health systems were very heavily hit. As you know, they were suddenly overwhelmed with all of these patients that showed up and had to be taken care of. They have to go to the ICU. They have to be put on breathing machines. Much of the research that was happening before was basically paused because patients were afraid to go to these health systems unless they were very sick because they knew that if they go there, they're probably going to be surrounded by other sick people. There was no vaccine yet. At some point in May, 2020, there were some reports that show that clinical research in the US had dropped by 90%. It's obliterated. That is obviously a supply chain shock that is hard to overcome. For a while, we were at a standstill. We had been on a roll. We had been growing rapidly. We closed a hospital deal every month of the year, January, February, March. And suddenly we were flat until probably November when there were talks of vaccines. This was to become the new normal; people were coping with it. And business started to pick up again. That initial stand still was worrisome, but we felt like we had a good team. We knew what to do. We kept calm and moved on. And in 2021, the opposite was happening. Because of the success of getting these vaccines to clinical trials so quickly, people started to think: "Well, the old ways of doing this where we expect to have to wait two, three, four years before we can take a new therapeutic to clinical trials. We've seen that we can actually do this in months. Can we also revisit this for our other clinical trials, not just COVID?" We're seeing a new impetus and a new opening of the minds of many of our stakeholders that think: "Let's try better ways of doing this."

Roland Siebelink: The exception became the new rule in a sense.

Wout Brusselaers: Exactly. It's really good. And sometimes you need that. The paradigm has to shift a little bit for people to become more open-minded and accept transformation. And in healthcare, that is a tough sell because healthcare organizations are by nature conservative. Their goal is not to rock the boat. They want to take care of patients. They don't take any risks. If something goes wrong with patients, it's bad. There's a lot of structural inertia you have to overcome before you can sell something new and something that's disruptive. And then selling them a software is one thing, asking them to change their workflows and go to what some people call a digital transformation, that is complete blasphemy. You have to work that into your process very slowly and carefully to make sure that you don't rock the boat too much and you still get to the results.

Roland Siebelink: Let's double click on that a little bit because that's definitely something that I see a lot of founders struggle with in targeting more conservative cultures, working with government, with health care, with some older industries. What have you all learned at Deep 6 AI about selling more effectively into those kinds of cultures?

Wout Brusselaers: It's a great question. And I think that alone could launch a conversation that could go on forever. It's important to really understand who you're dealing with and to have a lot of empathy for where they're coming from. Again, there's a lot of arrogance sometimes with founders. We know things better, we're going to change things, we're going to do this, et cetera. And you have to stop that. You cannot work with that. In sales terms, somebody once told me if you want to close a B2B enterprise deal, you have to earn the right to move to the other side of the table. And what they mean is instead of facing somebody, you have to earn the right to sit next to them and become a partner and somebody will help them solve problems. It's the same thing here. What we thought originally, we're going to sell software to our clients and that's it, has really evolved into a much more integrated model where the selling of the software is only the start of the journey. Once you do that, you have a customer success team that really helps navigate our champions to sell more internally and to actually help their internal people become more successful. In a hospital, or in a healthcare organization, you do not have the typical linear, top-down hierarchy of decision-making. If you're selling to a bank, if you're selling to a software company, you typically have a C-suite of people that make a decision in small groups and they can act quickly. A hospital is more like a conglomerate of different departments and therapeutic areas. You have an oncology center, you have a cardio center, you have a neuro center, you have a surgery center. All of those people have their own leadership, both administrative and clinical. They have their own tools, their own systems, their own budgets, and they're tenuously kept together with a vague C-suite that makes some decisions but has very little to say. In a way, once you land a deal with one of those stakeholders, you have to make sure that you make them very successful. And then you use them as agents and advocates to go from oncology to cardio, from cardio to neuro, et cetera. That has become a significant part of our organization.

Roland Siebelink: As you say, as a founder, you often have to overcome a certain arrogance of knowing better and starting to really meet the customers where they are.

Wout Brusselaers: Absolutely. There is a historical contingency of why they are the way they are today and what they're doing. You have to understand what they really care about is the safety and the well-being of their patients. And if you try to disrupt that and say, "Well, try this now and do this," and you don't keep that mindset first and foremost, then you're going to fail because there's just a clash of cultures.

Roland Siebelink: Absolutely. Very good. Sorry for the distracting question with the pandemic. But yes, I do have to ask in a business like this. What are some of the broader trends that you've been either fighting against or tapping into with Deep 6?

Wout Brusselaers: It depends. There are some specific trends that you're seeing. I mentioned datafication. I mentioned patient centricity, precision medicine, et cetera. You have to have COVID. There's the whole economic model of healthcare in the US, which is very much a fee-for-service model where you actually pay for better outcomes, which is strange, or you pay for more healthcare, which is why it's such a big portion of our GDP as well. And hopefully, and luckily, we're seeing some changes there with the shift to more fee-for-value. And that's a trend that we keep in mind as well. There's also the trend of increased patient diversity, which goes hand-in-hand with patient centricity. Patients are more and more active advocates for their own health. They do research, they have questions, they come to doctors with some information that they've already done. They use this tool called Google, which makes them pretty knowledgeable, for better or for worse. You have to take that into account as well. You have to make sure that you deal with patients and you invite them to be part of the conversation. That is one of the things. Outside of what I call those clinical trends that affect how we do business, there are other things like the whole investment climate has changed recently. We flipped the switch from the end of 2021. It's crazy evaluations. Everything now, the fear of inflation, and a slow down and other things. The Ukraine factor that we have in play now. But also the war for talent. With that inflation and with working remotely. Those are things I think about a lot. We are now de facto fully remote. We do quarterly on-sites where we bring the whole team together. We've been doing that since the end of 2020, so very early on. I think it's very important to celebrate the culture, make sure you know each other. We've doubled in size, so half of the people on the team today, I've barely met or I've seen them only in Zoom rooms. How do you navigate that? And is that something that we want to stick with forever? Today, talent has the strongest hand in negotiations. You meet the people where they are because it's truly a war for talent. We hire people where we can, which means we are decentralized. But I am thinking about a lot - is this truly optimal for us but also for people because there is a human element? Even if I spend eight hours a day talking to you on Zoom, it's not the same thing as being in the same office. I say, "Well, let's grab a coffee, let's do this, let's just talk about other stuff." It constrains the conversation. It makes it much more utilitarian than it normally would be. How do you deal with that? I wish I had the answers; we don't. But we're trying to take it into account. We don't want to request more than people are willing to give in terms of coming back to the office. We want to make sure that we do what works well for them. But we also want to make sure that people don't suffer from that inadvertently by being more isolated. And we also don't want to have an army of mercenaries working for us, where they're working from home and I can change my logo or my LinkedIn profile and I'll do the same work somewhere else. There has to be more skin in the game than that. We really care about our values and about our culture. We try to find people that are equally passionate and foolish and obsessed with doing and solving smart problems as we are. And it's fun to have those people together. And if you only work together but you don't have those other things where you connect and interact, it's a shallower way of working.

Roland Siebelink: You're already saying we're hiring people where they are, and you're mostly remote at the moment or de facto fully remote, does that mean you hire all around the world or all around the US or California?

Wout Brusselaers: At this point, still in the US. I would be open and we're definitely thinking about growing globally eventually because clinical trials are global, so we have to be where those trials are. One of the biggest things is that you have to know your limits. I'm sure you speak some German, so you have to focus on the things that you can master, that you can do to grow. Taking up too much, too soon is the death knell for a company like us. What that means is that in the next two years, we'll likely try to grow in the US. And the goals are that we will only hire in the US because we deal with sensitive clinical data and that data can not cross borders. We don't want to have people outside of the country to be able to deal with that, even though all of our data is secure and de-identified, we just don't want to take any risk there.

Roland Siebelink: You also mentioned how important the values and the principles of the company are. Can you talk a little bit about what is the ideal Deep 6 employee in terms of the values and the people you like to work with and who are the kind of people who need not apply.

Wout Brusselaers: Yeah, that's a great question. And it's an important question. It's perhaps the most important question. It's how you build a team and who you work well with and who you don't work well with. I think one of the biggest rewards you get out of a professional career is working with like-minded, smart, passionate people. All the rest is gravy. Of course, you want to be able to pay the bills. You want to do a lot of stuff. But working with smart, passionate people, that's the ultimate reward. We're looking for people who are very much mission-driven, who like what they do, are excited about what they do, are passionate, and want to do that. We're looking for missionaries not mercenaries in that sense. The next thing is one of our core values is we act on current best thinking. And that means a lot of things for us. But what it really means is we like people who can learn fast, who can gather data, make decisions, act on the decisions, and then say a couple of weeks later, "We have new data, let's change things. We learned something from this. It wasn't entirely right, it wasn't entirely wrong, but now we can tweak." That overlaps somewhat with the agile methodology that has many other downstream effects. One is the fact that you do your best to get good data, make data-driven decisions. But two, you're not looking for perfection. The moment that you feel like you have a good hypothesis, you test it, and you allow your team and yourself to fail and fail fast. And that is a really important cultural trait that I feel many people that come from bigger companies often lack. They're afraid to fail, and they're afraid to admit that they could be failing. And by the time that you actually show that they're failing, it's too late and you're no longer failing fast; you failed a long time ago but it's slowly come to the surface.

Roland Siebelink: I love how you really cornered that as a team in acting on current best thinking.What I see a lot, to your point, is that either people optimize for what they say, "We need the data-driven culture," which is true, but then can often mean it's an excuse for "Let's wait until we have a hundred thousand data points and we cannot take decisions for the next three years," or it's more like, "Okay, we need to make decisions fast and fail fast." But then that means you can just go on intuition without any data points, which is also wrong. And I love how you freely captured the best of both worlds in that current best thinking, allowing for change of mind when new data arrives.

Wout Brusselaers: Exactly. It's a fine line. I'm always flabbergasted when in US politics, people are being insulted for flopping when they change their opinion. I think it's a sign of a thinking person that you change your opinion based on better data. Why would you be afraid of that?

Roland Siebelink: Yes. Or a sign of a completely dysfunctional process. But as you said, that could be another three hours of conversation for sure. Okay. Very good. How big could this become, Deep 6? How big do you see this? What is the big, hairy, audacious goal? The huge mission behind this? Where do you want to be 10 years from now, Wout?

Wout Brusselaers: It's another great question. And again, I try to avoid having too much hubris or whatever, but we think this is a huge opportunity. What we want to be eventually - all the best clinical trials run on Deep 6. We want to be the platform that all clinical trials all across the world run on. And that means we want to connect all the patients with all the researchers with all the caregivers with all the sponsors, eventually on a single platform that drives all those decisions seamlessly and very streamlined and faster. And our goal is honestly at that point to become transcendental. We can almost move into the background where we just build a platform and we leave it to all of those stakeholders to interact and we provide all of the piping for that.

Roland Siebelink: The piping and the plumbing, absolutely. I love to hear that. I think that is mostly what I wanted to cover other than my last two questions. The first of the two last ones is always, after all this learning you've done on your entrepreneurial journey, what is the one piece of advice you would give to founders coming after you?

Wout Brusselaers: It's hard to distill it into a few maxims that are of general use. As I said, focus on learning. Turn everything you do into learning. Try to always have a plan. And what I always say is focus matters, ownership matters, talent matters, culture matters. And it goes from there. If you really focus on what you want to do, you can focus on a couple of key points or hypotheses that you want to test. You can turn those into some quick experiments and then you can learn from those and build it and test the right. If you do that well, what is really important next is ownership. I have this own framework that I've invented, which I call owner-builder-leader. I'm always looking for people who are an owner. That means I ask you to do X, you will deliver X, you will communicate about X. If you're running late, you will tell me. If not I can just say: "Come back when we're done." You deliver on spec and on time. That's what I'm looking for in people, That ownership. Ideally, the next step for these people is to start being builders apart from owners. I'm building systems and a framework in place to make what I've just done more scalable. And I can not actually extrapolate that across the organization. And then ideally, those builders that have built teams and tools and systems, et cetera, they become leaders. And I think about what else can we do? We've been doing this and we built this, let's think about what lies at the other side of the horizon. And for that, you need to focus, focus leads to ownership, ownership leads to talent. You need to have people that can do that, that are strong owners, who can become strong builders and strong leaders. And to attract those, you need to have a strong culture. Focus matters, ownership matters, talent matters, culture matters. And as you find the bridges between all of those, I think you have a strong foundation for success.

Roland Siebelink: Okay. I think that is actually distilled really well, much better than anticipated in answering this question, so very well done. And the last question is always more tactical. If people want to learn more about Deep 6, where should they go? What should they download and how can they help you most?

Wout Brusselaers: We are a B2B enterprise organization. We don't have a simple app to download. But I'd say start by going to our website. We love to hear from people. We love every feedback. We're a growing team. We're looking for talent. We're looking for those owner-builder-leaders. We want to encourage people to participate in clinical trials because it does help save lives. You help bring life-saving cures to patients faster. Either reach out to us and we can point you in that way, either because you want to work with us or want to know more about clinical trials and what you can do, or go to, which is a government website where you can look at clinical trials. Related to Deep 6 or in terms of helping clinical trials in general, there's a lot that people can do.

Roland Siebelink: Okay, really appreciate it. Thank you so much for your time and very insightful answers today, Wout Brusselaers, the CEO and founder of Deep 6 AI, joining us from LA today.

Wout Brusselaers: Thanks, Roland. It was a pleasure.

Roland Siebelink: It was a pleasure for us. Thank you so much.

Roland Siebelink talks all things tech startup and bring you interviews with tech cofounders across the world.