She quit her job without a startup idea. Then built a $550M startup from a bad dentist appointment. I Wardah Inam, Founder of Overjet

Wardah was a PhD graduate working at a biomedical imaging startup. But all it took was two different dentists giving her two totally different diagnoses for her to quit her full-time job.

She was unemployed with no idea what to build.

All she knew was that something was broken—and she had to fix it. Her startup Overjet is now the #1 dental AI platform, valued at $550M.

Why you should listen:
- Why founder obsession is a key trait
- Why learning speed is the main KPI early on
- Why staying close to customers is how you build the right product.
- How to get customers to share data with you and let you spend time in their office, watching how they work.

Keywords
dental AI, startup journey, funding challenges, product-market fit, impact on dental industry, speed in startups, biases in fundraising


Timestamps:
(00:00:00) Intro
(00:00:52) The Start of Overjet
(00:05:05) Obsession Increases your Chances
(00:08:15)A Bad Dentist Appointment
(00:11:35) Surpassing Dentist's Analysis Using AI
(00:15:52) Finding a Co-Founder and Fundraising
(00:24:22)Finding True Product Market Fit
(00:27:33) One Piece of Advice

Transcript

Wardah: 0:00

For me, what I learned over time was that speed is probably the most important thing in a startup.

Pablo: 0:07

Welcome to the Product Market Fit Show, brought to you by Mistral, a seed-stage firm based in Canada. I’m Pablo. I’m a founder turned VC. My goal is to help early stage founders like you find product market fit. Wardah, welcome to the show.

Wardah: 0:24

Hey, Pablo, thank you for having me.

Pablo: 0:25

I’m looking up Overjet online, and it’s like building the future of dentistry. It’s like dental AI, which if you just come at it from today’s landscape, you’d think, oh, it’s just jumping on that AI bandwagon. But you’ve been doing this for over six years now. You started in 2018. Frankly, as a VC, I don’t see a lot of plays in dentistry, so I’m really curious to hear how this started and how you ended up here in the first place.

Wardah: 0:51

Going back a little, I did my PhD in clean energy. I was very interested in making the world better. Then I started to realize that in clean energy, a lot had to do with policy and less about the technology. I think we had a lot of good technology, but what was stopping us was the policy, and that’s something that I felt that I could not provide the most value in. I basically moved towards healthcare and started making conscious effort to – so I did a postdoctoral fellowship working on biomedical sensing using machine learning. That was a conscious effort that I made, and that got me more towards this space. After that, I was working on a startup doing biomedical imaging. At that time, I changed my dentist and got a treatment plan, which was very different than I had received before. That’s how I started getting into dentistry. It’s a fascinating field.

Pablo: 1:57

What were you doing? Just tell me a bit more about the last startup. You’re doing biomedical imaging. Just tell me more about what that is.

Wardah: 2:02

We’re working on MRI data, making that faster to acquire, more quantitative, so that we can get focused on more preventive care and being able to catch disease faster so you don’t have to suffer from the disease. That was the thesis behind it, and we were building out MRI technology for that.

Pablo: 2:21

Were you at the time thinking, I’m doing this to learn about startups so one day I can start my own? Or was that not really a thought in your mind?

Wardah: 2:29

All through my PhD, I was building startups where some of them worked. Some of them did not work. I was commercializing anything I built. I think that was very inherent in what I was doing. Even in my postdoc, the project I was on was also a startup that was being built. I did join Q bio because it had a founder who was a successful founder. His last company was a firm, a successful company as well. It was about learning as well, but it was also I fell in love with the mission of the company. I didn’t want to join the company. I got introduced to the founder and he convinced me that it was an amazing opportunity and to really have an impact. That’s why I joined.

Pablo: 3:12

You had this itch to maybe start something. You’d already started a few things in the past. Then you have this experience with a few dentists. Can you expand on that? You were sharing that story with me earlier.

Wardah: 3:22

My personal experience was that I moved locations, so I changed my dentist when I went from one neighborhood to another neighborhood. When I changed my dentist, the new dentist gave me a treatment plan, which was very different than I had received before. I was surprised at the fact that why there was this much discrepancy. At that time, I asked the dentist to provide me with the x-rays. I started reading more about dental x-rays, dental one-on-one.

Pablo: 3:46

You’re chasing what at this point? You’re literally trying to understand your treatment plan, or you’re already thinking maybe there’s some sort of – because you know about imaging?

Wardah: 3:55

I’m just trying to understand my treatment plan and it just didn’t make sense to me. There was something off. For me, it’s one of those things. I was just trying to understand how important was it, why was there this discrepancy, the fact that I didn’t know anything, what was important to me as well. I started just reading up online, et cetera. Then I shared my x-rays with other dentists as well and I got different opinions. That what didn’t sit well with me, where it felt more of an art rather than a science. There has to be one truth. What is it?

Pablo: 4:27

This is on your free time. You’re working full-time. You’re doing what you’re doing. Then in evenings and weekends, this is your passion project?

Wardah: 4:35

Yeah, and it didn’t start as a project as well. I started getting more into it, and then it got to a point where I said, okay, I’m spending more time on this than I’m spending on my day job. Anything I do, I do it fully in.

Pablo: 4:47

Maybe let’s tangent on that a little bit because work-life balance gets a lot of attention, let’s say these days, but it doesn’t really fit, does it, with the zero to one? You talk about obsession and it’s a trait you see I think in a lot of successful founders. You think that’s a must have, you need to be obsessed to make things happen in those early days?

Wardah: 5:05

I think one is about my personality as well, where I’ve found that for me, I will not be happy if I force myself to do the 9 to 5. That’s not just me. It’s one of those things that is believe that’s better, but the person I am, I really enjoy the things I do. Telling me not to do that is you’re taking that joy away from me. It’s not the positive. For me, that balance is different as well, and be having the freedom to work on things that are exciting, et cetera, that I want to work on and solve difficult problems. That’s what gets me excited. To your point of whether it’s necessary, I think everybody is different just in general, but I don’t think there is any shortcut to hard work. I think hard work is real and it actually gives you an edge. All my life, it’s something that I’ve believed in very heavily, and it’s something that it might – again, there might be people. The people I know who’ve been successful, whether they were professors doing their thing, these were some of the smartest people. I don’t think they needed to work the way they were working, but there were some of the hardest people at the highest level of their career that they could be. Nobody was telling them to work this hard, but that’s I think was a need to them. I think similarly in startups as well, I don’t know and there might be. I don’t know any founder that I admire that is not putting in their all. It is hard. It’s not other easy things to do. But if you’re solving a very hard problem, there might be that you got very lucky and there’s PMF from the beginning, and it could work out. I’m not saying it has to be always true, but I think you do absolutely increase your chances of success if you are putting in everything.

Pablo: 6:56

I think that’s true. It’s a competitive landscape and it’s really not like there’s skill, and then there’s hard work, and there’s going to be people that do both. Obviously, if you’re not one of those people, your odds of success are probably lower than somebody that’s potentially competing with you and is equally skillful, but also just putting in a lot more work. There’s balances all these things.

Wardah: 7:14

Pablo, I feel especially in the beginning, it’s all about learning. Again, there are no shortcuts to learning as well. You need to learn and you need to put yourself in different experiences. You need to understand different viewpoints. You need to understand your customers, and you have to have conversations. That’s just time that it takes to learn everything new. The faster you learn, the faster you execute, the faster you can make progress as well and learn from the mistakes, because everybody’s going to make mistakes. I think it’s almost inherit to – again, you could get very lucky with everything just working, but most things just don’t work. If they were that easy, then it would have been done before.

Pablo: 7:57

Back to the storyline, you’re obsessed about dentistry. You’ve unlocked some potential, let’s say problems, which is that it’s more art than science. But do you have an actual startup idea at this point when you start thinking about, okay, maybe I need to get out of my day job because I’m just spending so much time over here?

Wardah: 8:15

This is one I don’t recommend, which is I did not. I had a thesis, and my thesis was that if I did not do any – just to give you a sense. When I was understanding the field more, what I realized was that there was a challenge in terms of how dental data was being analyzed, how diagnosis was happening. The way the industry was with private equity entrance into the space, say 10, 15 years ago, with a lot of money coming into the space with consolidation happening, and not having ways of measuring clinical quality. There were ways of measuring financial performance. There was no way of measuring clinical quality in dental practices. I felt that if we did not do anything about it, we would head in a very wrong direction. This was the time that dental service organizations with this consolidation that was happening, I think everybody realized this was the future. This was going to continue. Earlier, there was a lot of resistance to the consolidation that was happening. But then I think at the time that I started looking into it, I think people started realizing, because of the economics, this is going to happen more and more. If this is the trend that we’re going to go on, these groups do need a way to measure clinical quality. If they cannot measure clinical performance and clinical quality, they cannot improve it. That’s where Overjet started, or at least the idea of Overjet started to form in my head. It was almost I had to do something about it, so I quit my job and started shadowing dentists. There was not a name for it. There was no formation that happened at that time. It was literally for me to understand more of the space, and I had to go and actually see what was happening in dental practices.

Pablo: 10:00

How did you do that? How did you let dentists let you into their practice just so you could watch?

Wardah: 10:05

I did not have a good network there. I was trying to find dentists first. I think the first dentist connected with was in my building. The second dentist was literally, I was doing an Uber pool ride, and the person sitting next to me, I got talking. They asked me what I was doing. I told them I was exploring the dental space, and they introduced me to a dentist. That dentist ended up giving us our first tranche of data. I think it was more about just talking to anybody what I was working on and the problem that we were facing.

Pablo: 10:45

What was in it for them? Were they just being nice or was it just like – did you give something in exchange? They would get the alpha version if you ended up building something? How did you even convince them to do that?

Wardah: 10:55

No, I think they actually believed in the problem. I think for me, when I was stating the problem to them, I think they all saw the problem as the way I was seeing it. They didn’t know what the solution would look like, but they wanted to be part of the solving the problem. I will tell you this, the first time somebody shared that data with me, they actually said, no, thank you so much for working on this problem. That’s how they shared the information. It was a problem, I think, that people realized, and the fact that I was working on it.

Pablo: 11:27

What were some of the key things that you discovered in those early days through the work you were doing, through the research you were doing?

Wardah: 11:34

The way we started building this out was trying to understand the problem. We didn’t know that this could be done. We didn’t know what was possible and what was not possible. A lot of it was once we got the data in dental practices, we started just doing data analysis of and saying, okay, what is the prevalence of this disease? What does this look like? Why are people diagnosing a certain way? There was a lot of things that we found. For example, one of the things that’s key to Overjet even now is everybody thinks that there’s more overdiagnosis in dentistry than underdiagnosis. If I chat with most people, they will say, yes, there is an incentive for dentists to produce more. That is why telling you that there are conditions there, there’s the trust level suffers because of it. There is significant amount more underdiagnosis in dentistry and treatments that are not happening than overdiagnosis. That we found by actually analyzing the data and looking at the rates that CDC was saying in terms of prevalence of a certain disease, what we were finding, and how much it was getting diagnosed. I think there was a lot of interesting things that we found just from the data itself. Being in the chair, you can realize that dentist job might be tough. But if you are observing how they’re interacting with not only you, but other patients at the same time that you are in the practice, you start realizing how tough that job really is. Especially coming from the MRI side where radiologists are spending, a lot of time looking at x-rays and other MRI data, and usually in darker rooms, and with the right monitors, et cetera, that’s not what dentists are utilizing. They’re moving rooms. They’re changing lightning conditions. They’re having these old monitors that they’re making diagnosis on. There’s a lot of interesting learnings that I don’t think that the industry knew and had the belief in. But I think as we started to look at it from the data, and as a data problem, we started saying, the data quality is bad. The way things are being monitored is bad. The monitors that they have is bad. It’s not like dentists don’t want to do the right thing. The tools are not there, and we can do something about it.

Pablo: 13:54

What was that first solution that you conceived of? Because today, my understanding is you just automate that. You look at images and use AI to understand if there’s something wrong or make diagnoses. Was that the idea that you had at that point?

Wardah: 14:08

Yeah. From the beginning, we were looking at extras, and extracting or analyzing disease on extras. The interesting thing about dentistry is that if you went to a dental practice and you wanted to query how many patients have cavities, you won’t be able to query that because disease is not recorded in dental practices. Treatments are recorded, but disease is not recorded. The disease actually sits in x-rays, sets in notes, et cetera. That’s what we started doing was like, okay, let’s start to analyze disease. But what was really unique about us was that we started to quantify disease. Rather than saying, hey, there’s a cavity, we started saying, hey, here’s a cavity which is 25% of your crown area of the tooth. That means you need this treatment. That is surpassing dentist capabilities as well, but also making very objective. Now, you can actually communicate, going back to my original problem, where two dentists are disagreeing with each other. Now, you can actually say, oh, I’m disagreeing because of this, and this person is saying because of Y, because now you can start to quantify disease and start to say, okay, it’s X percent in enamel, it’s indent, and this is what it looks like. I think that helps in much more objective decision-making, but also then extracting that data so you can actually do many interesting things with that information and try to solve this more societal problem as well.

Pablo: 15:35

The other part of the storyline that we’re not talking about here is at this point, you’re a solo founder, you’re a first time founder, and you need to build a pretty sophisticated system. Do you go out and fundraise? What’s your move as the idea starts to form? How do you think about actually executing against it?

Wardah: 15:52

First thing was, okay, I need a co-founder dentist. I tried to find somebody who would be my co-founder in starting the company. With that, I started reaching out to many different dentists or talking to them, trying to figure out who would that person be. There was a dentist that I found who was doing his residency at Harvard, common connection we had that who used to be at MIT before and was interested in entrepreneurship. He was there in different entrepreneurship events, so I reached out to him. His name was Alex, and we started working on this project, but he was still in school at that time. The idea was that he would quit school, this was his residency, and work on this full-time, or at least take a leave of absence. But Harvard did not give him a leave of absence to work on the project, so he was with us part-time and helping us provide the dental knowledge that was needed.

Pablo: 16:48

Got it. Then what do you do about fundraising? Do you apply to some accelerators? Do you start pitching angel investors?

Wardah: 16:54

I had worked for a year. That means I had some savings. At that time, I thought that was a lot of money because that was my first job. I basically start utilizing the savings I had. That’s when we hire Deepak as well, who worked with me at Q bio. Before, he was at Amazon. Also, my parents had given me – when I had graduated and married, they had given me a check for putting down payment on our house. We used that. That was our seed money there. Then I started talking to dentist there, and then there was another dentist who provided the person who gave us the initial data. He provided the second check, and that lasted us for about a year.

Pablo: 17:37

Okay. In that year, where did you think about where you had to get the company to in order to raise that next round?

Wardah: 17:46

If I go back, now I can – I think sometimes you tell the story better, looking back. But I think at that time, we knew that we needed some customer to get the funding. We started pitching to investors as well. It wasn’t the easiest thing to get funding. I did not have a network of investors. Now, I know a lot of investors. At that time, I knew nobody.

Pablo: 18:09

What were some of the responses that you were getting from the ones that did take the meeting and, let’s say passed? What was some of the feedback you got back then?

Wardah: 18:17

I think right now, people think AI is the hottest topic, but I don’t think it was at that time. From my side, it was very obvious that this is what we should be working on. However, I think there was, because inference has a cost, the way software traditionally has not had a cost –

Pablo: 18:37

There’s no unit cost. The marginal cost is zero for software. That’s right, yeah.

Wardah: 18:40

That that aspect of it I think we just did not like, basically in terms of what margins could be, et cetera. That was one piece of it. It didn’t have major successes that people could point to. There were some companies like WIZ.AI. I think the first generation of companies which could not – and especially healthcare could not make it work at a larger scale. They did solve a very good problem. However, adoption of that technology was not really happening at the scale that visas would like it to work. There was actually a market problem as also there was an AI problem. The adoption was happening at that time. I would say third, this is something that I didn’t want it to be true. I still avoided it to be true, but I do think there was a bias in terms of for me to be a woman starting this company as well. Even if I was hearing it, I was not trying to internalize it, which was where an investor – this was an angel investor. They have a fund, a pre-seed fund, who basically told me that, hey, I really like you. However, the chances of you getting funded next are so low that I don’t want to put money in because I think the statistics they shared with me was only 2% of female founders get funded. If you get a male co-founder, I’ll fund the company, or male CEO. It wasn’t about a co-founder had a male co-founder, but it was getting the male CEO in. I don’t think they were doing it from bad intentions or anything. They actually believed it. That’s one where I think there was a little bit of this pattern recognition that VCs do. I think that that’s one where it was hurting us a little as well. The amount of traction they wanted us to have at that time was much higher than normally I would expect a startup to get funded.

Pablo: 20:45

Pattern recognition is a funny thing because on one hand, it’s the tool that both species need to rely on in order to understand what things might work, but it’s almost baked into pattern recognition that what you end up forming are biases. Some of them are helpful because you’re like, this thing tends to work, but some of them are actually, they create these crazy blind spots. Obviously in your case when it comes to female founders, there’s so many other examples of spaces, areas where people say, oh, nothing in this area could be big because there’s been nothing that’s big, and all of a sudden, you miss the next one. That’s a pretty wild story and black and white bias there, and going through that. That’s hard to imagine. Take me through then, because at one point you do overcome this. You over time raised over $100 million and just looking here like your seed round in 2020 was pretty big. It was almost $8 million. What happened between the time where you raised from that that you had your savings, you had the angel, you had a few, let’s say $100,000 or so to that $8 million round? What did you accomplish in that time? Because that’s a big seed round.

Wardah: 21:46

The first institutional check that was written by E14 fund, so MIT’s fund, I think they took a chance on us, so it took a chance on me. I think that was very crucial in starting to make it into a real company. Before that, I think of it as more of a project that was happening. That was one thing that happened, as well as we got space in at Harvard Innovation Labs. We had a space and we had an institutional investor behind us, and then it was literally traction. We had to prove that this works. We had to get customers. On one side, we were trying to figure out how do we commercialize this technology, which can have so much impact. We found that insurance companies had the biggest need and urgent need in terms of – they were reviewing a lot of claims manually, and they were making the same decisions that dentists were making, so they had hired a team of dentists who were looking at this data manually and making these decisions. We knew that we could help solve some of the – at least automate all the administrative tasks that they were doing, checking for completeness of a claim, checking for, is this the right x-ray, the data, et cetera? Then we could actually start to look at helping them in clinical aspects of it as well, so medical necessity. We started observing that, trying to figure out what that would look like. At that time, I think from the investor’s belief was insurance is very slow, so you guys are going to die before you get a contract there. Now, I understand it. I now understand how slow they are. But at that time, I think naivety helped us a little because I said, seemed like they’re moving. It seems like there’s a need, so we started actually pursuing that. I would say once we built enough that we could show a demo, we were able to get insurance companies. Then by the time, you mentioned the seed route, we had signed Guardian and Delta Dental. These are large insurance companies.

Pablo: 23:52

Why do you think insurance companies were the better first customers than dentists themselves?

Wardah: 23:58

For dentists, we needed FDA clearance, and that itself is a longer process. We started the FDA process while we needed actual customers to get the funding, to build what we needed to build.

Pablo: 24:12

Perfect. That makes sense. We’ll stop it there. We’ll end on the two questions that we always end on. The first one is, when did you feel like you had true product market fit?

Wardah: 24:22

For me, I started feeling that this was real when we started signing these insurance companies. The way it worked was we actually got these two insurance companies that we piloted in. We showed that it actually works and it could work. After that, there was this inbound from a lot of other insurance companies doing RFPs to see if this works. We basically won every RFP that we participated in. That hadn’t translated into revenue yet, but for us, it was like, okay, this is real. This is going to be huge, and this is going to have a huge impact, and the customers were seeing it. If you actually talk to even a customer right now, they see the same thing that they were seeing at that day. This is magical. People jaws were dropping when they were seeing our demos, and I think it’s harder to – revenue came in later, but I think that earlier, the demos were – when we actually do our demo, I would say, one out of three would get a wow. Then after that, when people would actually deploy it, mostly they would say, oh, the results were much better than we even imagined in the beginning. I think that’s where, for me, it started.

Pablo: 25:35

What was the number one RI for these insurance companies? Was it not, let’s say being able to deny some claims, or was it just efficiency and just having fewer people on staff reviewing claims as they came in?

Wardah: 25:47

I the same problem that we were facing or I faced was you show data to two different dentists. They’re going to say different things. How do you know what is right? Now, what you don’t want is a dentist on the other side saying, I don’t think this is what it is. Then they’re going back and forth with each other, and then now bringing in – and these are visualizations that come into play. There’s quantification that we add into this data. Now, two people can at least have the same rulers. They can have the same information, and then they can actually talk to each other and be on the same page. I think that is what really changed in dentistry. I think for these insurance companies as well, that translated into higher consistency in decision-making. Now, the amount of training that they had to do of all these people – and then people could spend time on the real things rather than spending time on things that were super easy. Everybody knows, as I mentioned it as like administrative tasks. Anybody can do that, and getting that out of the mix as well, and making sure that then they’re spending time on the right things. They’re focused on the right problems. Then even for the reviewers who were using this technology, some of it got fully automated, and that was great. These were all the sets, so now you’re making faster decisions, so it’s better for the provider. The ROI was pretty high in terms of financially, but also quality-wise, and provider experience-wise. It was a game-changer in how they were actually operating.

Pablo: 27:23

Last question, if you could go back to 2018 when you were just starting this company with one piece of advice for yourself based on what you’ve learned over the last six years, what might that be?

Wardah: 27:33

For me, what I learned over time was that speed is probably the most important thing in a startup. Inherently, I am very impatient, want to do things fast. That’s something that I have, but I’ve always thought of it as almost a negative in me. I’ve always been like, oh, I’m very impatient. I’ll use the word impatient rather than like, oh, I need to go fast. There are things that I realized a little later in my career that they were actually the positives and I need to actually instill the same culture in the company rather than saying, hey, this is the negative, and I need to slow down, and I need to do things a certain way. I think that’s something that I would change if going back.

Pablo: 28:22

Perfect. We’ll stop it there. Wardah, thank you so much for spending the time.

Wardah: 28:25

Awesome. Thank you, Pablo. Thank you so much for speaking with me.

Pablo: 28:30

I just gave you content that you liked so much. You actually listened to the end. Guess what, you didn’t pay a single dollar. Not only that, I didn’t even put any ads in your face, so you just got a bunch of content for free. Now that I’ve delivered that value, I’m asking for something in return. Open your app, open Apple Podcast, open Spotify, open whatever app you use to listen to this and hit that follow button. It’s actually going to help you because it’s going to help you make sure you don’t miss out on the next episode, which you like so much that you listen to the whole thing.