Emerging AI technologies are set to transform dentistry—and this week’s guest is at the vanguard of the seachange.
Ophir Tanz says artificial intelligence won’t replace dentists but will support and free them to focus on patients and clinical time.
His latest project, Pearl, is already providing AI diagnostic support to a profession that carries out more radiography than any other medical discipline.
Ophir explores the potential use cases for AI in dentistry, shares thoughts on the regulatory landscape and gives the lowdown on the challenges of training diagnostic AI in an industry with a surprisingly inconsistent approach to identifying and treating pathologies.
In This Episode
02:05 – AI, GumGum and Pearl
09.50 – Use cases: data and pathologies
13.51 – Training AI
17.13 – Neural networks
21.10 – Patient communication and case acceptance
22.31 – Regulatory hurdles
24.14 – Early detection and diagnosis
26.50 – Future applications
29.46 – Pricing and distribution models
32.54 – Ophir’s story
34.48 – Drive and five-year horizon
38.04 – Superpowers
42.09 – Blackbox thinking
50.13 – Competition
52.39 – Highs and lows
56.58 – Business inspirations
57.57 – Last days and legacy
01.03.02 – Fantasy dinner party
About Ophir Tanz
Ophir Tanz is the founder of Pearl—a provider of AI-powered diagnosis and practice performance software for dentists.
He previously founded successful interactive and AI brands in media, branding and mobile spaces.
It’s the hardest thing to do. I feel that I have done that many times. Many times, Yeah. I mean, over, you know, a thousand plus employees and you have a lot of people that are very dedicated and their heart’s in the right place, but they’re just the wrong fit for your organisation for any number of reasons, either their skill set thing or it’s the talent thing or it’s a culture thing, right? And the best thing that you can do, in my opinion, is be honest, be kind, be generous with those people and try to be helpful to them. But you have to protect the enterprise as a whole. And you have to make, in my opinion, very swift decisions as it relates to the to the personnel at the organisation. And this is why I don’t like the family analogy for companies, because you can’t fire your family. You know, I’ve tried but, but you know, a professional sports team, which is a good analogy that you know Reed Hastings uses a Netflix I’m kind of stealing that but I think it’s the perfect one because we’re recruiting top athletes and we’re expecting a lot from them and they get released if they’re not contributing in the way that we need them to. And that’s a much better analogy. That’s much more accurate, I think.
This is Dental Leaders the podcast where you get to go one on one with emerging leaders in dentistry. Your hosts, Payman, Langroudi and Prav Solanki.
It gives me great pleasure to welcome Tanz onto the podcast. As a serial entrepreneur who’s lately turned his hand to dentistry with Pearl AI, which has actually spun out of another computer vision, AI company called Gumgum. Computer Vision. I mean, it sounds like a crazy idea, but getting meaningful information out of digital images, I guess. Thanks for coming on the pod. Lovely to have you.
Yeah, it’s great to be here. Thank you for having me.
Sophia, give us the sort of the lowdown on someone who, you know, we were just talking of of her as someone who’s had so many successful companies, You know, your your previous one, the Gum Gum company was looking at sort of logos. Is that right? Tell us what that company was doing that led you to to to Pearl?
We did a number of things with computer vision. But generally what we did was crawled much of the known Internet every day and on an ongoing basis and developed an understanding of sort of non-textual media. So primarily imagery and video and eventually incorporated text as well, and sort of made that computer vision analysis of that non-textual data available for a variety of purposes. And the reason that’s important is because historically that’s all kind of black space as far as the Internet is concerned, because nobody’s been really able to interpret it. But it obviously contains a lot of very important information. And that’s kind of what Gumgum was focussed on doing, was illuminating that part of the Internet.
So but what were the applications?
So there were a number of applications, so advertising, advertising, targeting work that we did around sort of valuing brands and their exposure online and on televised sports events, stuff like that. Basically anywhere where a brand or a company is interacting in a visual visual medium, we were able to sort of help. So what we did specifically with sports sponsorship, for example, is we analysed every moment that a sponsorship came into view and actually quantified the value of that moment of exposure and that became a currency for brands. And Rightsholders also worked with nearly 100% of Fortune 100 brands in relation to their advertising initiatives online as you related to both targeting but also providing access to our proprietary ad formats.
And then you sort of internally incubated Pearl out of that technology.
Yeah, we internally incubated a number of different computer vision driven applications as a function of that technology. So really around 2010 or so, what you’re able to do with I started to get very, very powerful and interesting, really a dramatic shift from what was possible before. And that’s really because of the resurgence of neural networks, which are an old technology that came back into fashion and suddenly were very plausible because we had the computing power to sort of run these things efficiently. And all the craze that we’re currently seeing around AI, everything from Chat ChatGPT on down is really a function of that shift and that resurgence of this particular branch of artificial intelligence. So it became very clear to me that while we were applying this technology, you know, it’s a very successful effect at Gumgum, there were other applications that were also very interesting, and we launched a number of initiatives internally, one of which was a personal passion project of mine, which was applying this technology in the field of health care. And to radiography. I believed at the time, and I think many people increasingly believe that AI is going to play a very fundamental role in becoming a standard of care, in elevating the quality and precision of diagnoses across all forms of radiography.
So then what led you to Dental?
Yeah, so Dental is an interesting one. You know, my father is a retired dentist, so I did kind of grow up in that environment and it was very familiar to me. But we did initiate a pretty systematic review of various forms of medicine and did look at a variety of opportunities outside of dental as well. The reason that we landed on dental in the end was, you know, I obviously was already interested in it. I’m just having some familiarity. But I think more importantly is if you look at all of medicine, more radiography is captured in dentistry than any other form of medicine. You have access to data. A little bit more readily because it’s not quite as sensitive as other forms of radiography. So you have things like brain cancer and various other forms of cancer, and there’s a lot of difficulty in getting access to that data and it becomes very difficult and costly and ultimately makes it more difficult to train effectively. If you also look at the the nature of dentistry, we liked that we were not competitive with the actual customer. So dentists by and large need to perform radiographic review in order to do their job, but they’re not first and foremost, you know, priding themselves on that fact.
And that’s not the totality of their value in the world. Is it something that they have to do? So the idea of unburdening practitioners so they can more effectively diagnose, which is the first step, often in sort of, you know, identifying the status of oral health and ascribing a treatment plan to us was attractive because if you look at radiology, we and the integration of these types of tool sets there, there is a lot of concern and consternation and resistance. And ultimately I do think that these technologies do work best with both humans and machines working together. But in dentistry you just don’t have that that high, high level of sensitivity. And then those other elements on the business side in terms of dentistry being much more fragmented and entrepreneurial, you’re not selling into very large hospital groups and, you know, stuff like that. So for a variety of reasons, we thought dentistry was just a wonderful place to apply this technology. And we also believe that because of the characteristics that I mentioned, it would very likely leapfrog the rest of medicine in relation to becoming a standard of care. And I think that’s actually playing out actively today.
What stage are you at with it? Have you got regulatory approval across different different areas? Are you confident it’s better than the human on its own?
We’re actually at a much more fun and exciting stage. The past few years have been about red tape, regulatory approvals, clinical trials, you know, things like GDPR compliance, just a lot of the, you know, IP sort of oriented work as well, a lot of R&D. And those are the necessary things that you need to engage in in order to to do this properly. Where we’re at now is we have regulatory clearance across over 100 countries. We have engaged in obviously successful FDA clinical trials that very rigorously prove efficacy. We have distributed infrastructure globally, so we’re now operating in the UK, in the EU, in the Middle East, in Australia, in Canada and America and other territories. And now it’s really about distribution and integration. So the company is growing very rapidly, the technology is becoming adopted very rapidly and we really do feel like in the next few short years it will become a fundamental standard of care. So this really is in the hands of many practitioners in the operatory patient facing worldwide today.
For those who don’t know what Pearl is or don’t understand how AI works in this, could you just describe to the average dental practitioner, you know what this is from? From my limited understanding, I see it as that now. Instead of me looking at a x ray and trying to figure out is do I see a bit of decay here or do I see a bit of issue here that you can fast track me to identifying stuff that maybe my eye would not have picked up? And the more times that this technology is used, as time goes on, it becomes more accurate and more efficient and better at picking up those things, making me essentially redundant. Looking at radiographs eventually is is my interpretation close or what would you say is.
Yeah, I mean, I would start off with some data. So, you know, we ran a study with a non-profit that we helped found called the Dental Council, which is intended to study the intersection of dentistry and AI. And in one of our initial studies, what we did was we asked 136 practitioners to diagnose and treatment plan a single patients. And what you found in that study was really striking. There was almost never greater than 50% concurrence. If you looked at select teeth in the mouth, you know, it was basically a coin flip as to whether or not there was decay present. I remember two three in particular in the study. I could share it with you. It’s available on our website as well. You know, it was like 51%, you know, decay, 49%, no decay, you know, 60. Like it was a. 60 over 40 on recurrent decay, and the treatment costs came in between $300 and $36,000 for the same patient. Similarly, there was another study back in 1997. A journalist working for Reader’s Digest decided that he wanted to do a review of the state of dentistry in America, and he travelled to 50 states and got 50 diagnoses. And I think his results were something like between 0 and $40,000 in treatment cost recommendation. So there’s this massive lack of consistency. We work a lot with university partners and all the deans of all the schools of dentistry say the same thing, which is that they know that they’re not sending dentists into the field proficient at identifying a range of pathology, and they’re expected to learn on the job.
But that’s very inconsistent and there’s not really a good feedback loop there. So the need is obviously very real. If you look at our clinical trials, we were able to show that we surfaced 37% more disease on average per radiograph encountered. So this is everything from various types of caries. Obviously, interproximal are often missed, but you have things like calculus that are often missed or early bone loss or periapical lesions or margin discrepancies. The list goes on. Our technology is called second opinion, so we are a real time patient facing non human in the loop tool that is simply highlighting radiographs and where it’s relevant. Also sort of measuring areas of decay or bone levels or other sort of pathologic and non pathologic conditions that are present. And the idea is to point out those areas of interest to a practitioner so that they can make their own assessment, they can check things clinically and they can ultimately then begin engaging on a treatment plan. So at our core, the thing that we’re most known for is this tool called second opinion, which is the most advanced and the most comprehensively regulatorily cleared product on the market as it relates to this kind of diagnostic AI. We have other tools as well for practitioners that go even a level deeper that interact with the practice management system. I’m happy to talk about that, but I’ll just stop there and see if see if that’s clear.
How do you train a computer vision model and how bad was it originally? I mean, someone I’m always interested in people who I mean, you are a veteran of this, right? I mean, there can’t be many people who were in on this in 2008. How bad was it before and how quickly has it moved on? And how do you train it? Do you show it loads and loads of radiographs and humans checking up its work.
Painfully and with a lot of money is the truth. So what do you do?
That sounds that sounds like the regulatory. I want to ask you about how much that cost you.
Everything costs a lot. This is not a cheap thing to do. Well, yeah. So we’ve raised, you know, about $31 million to date and are, you know, in Q1 likely engaging in a much, much larger raise on top of that. It’s an expensive endeavour. On the training side, what you need to do is, you know, a massive quantity of representative data. So you need to have represented demographics and geographies and sensor types. You want to do things like digital radiography and phosphor plates and just make sure that it kind of covers the market generally because this technology is intended to be used in market with any sensor and any imagery that’s thrown at it. And then you very painstakingly with hundreds of dentists initially, and then you whittle it down in various ways, label that data. So you go in and you annotate it and you sort of mark up where there’s disease and you have methodologies that are statistically valid for for litigating disputes between sort of two different practitioners or three different practitioners. And that’s actually one of the challenges with with the FDA, for example, is that they looked at the data and they said, we don’t know how you can have a ground truth set that you’re going to test yourself against because nobody’s consistent to begin with. So we had to engage in a lot of statistical analysis to actually show that we were able to, over time, develop a ground truth set that was actually valid and reliable.
And then you start to train. So you start feeding it into the neural networks and you see what comes out. But what people are often not familiar with is the heavy amount of pre-processing and post-processing and heuristic rule sets that are then layered on top of that. You know, certain things might be detected by the AI that just would never make sense, right? And you can kind of eliminate those and you kind of go through that process. So, you know, is really refined on an ongoing basis. We put out models every couple of weeks. It’s always reinforced with new data when we run something. Thing through our system. You’re really running through over 30 neural networks, not just one massive one. And we have, you know, 8 or 9 neural networks that are specifically focussed on different types of carries. So as far as the machine is concerned, those are all different models. And you know, a similar philosophy applies to other detections. So when you’re dealing with a medical application with a very high bar of expectation and requirement, then you actually have to go well above and beyond what you would do in other computer vision oriented applications. Although I will say that the general methodology, whether you’re training on cars or cats or or dental pathology, is pretty much the same process.
What is a neural network? Sorry for my lack of thingy, but you mentioned that, you know, 30 different neural networks and in one particular area, say eight different neural networks. What does that actually mean?
Yeah, good question. So a neural network is a branch of artificial intelligence, right? There are various approaches to AI that have existed over the years, and a lot of them get a lot of excitement and most of them have generally kind of fallen flat of expectation. What the neural network approach is is effectively a somewhat rough but also accurate representation of the human brain. So you have synapsis synapses and axons and you have this sort of network of synapses effectively interacting with each other in the brain and based on the understanding of the human brain. Actually in the 50s and 60s, this notion of a neural network was conceived of this notion that if you just feed raw data into this network, it will start to figure out patterns in that data and it’ll start to reinforce itself toward the right answer. That’s why this is often called reinforcement learning. So basically what you’re doing is teaching a computer to understand things conceptually versus describing them. So if you think about, you know, an apple, if I tried to with traditional programming describe an apple, I would say it has this curvature and it has a little thing that looks like a twig and that has a certain colour and a curvature and it could be these colours and, you know, that’s all fine and good, but then you feed an image of a rotting apple on the ground or an image of a monkey holding an apple or an image of an apple and slices and suddenly that thing doesn’t work well at all.
So that’s just not a good approach to trying to have a machine identify where an apple is or is not. However, with a neural network, you’re able to feed in every example of an apple that you could think of, and it will start to identify the edges and the colours. And ultimately the concept of this notion of Apple in much the same, you know, I have a two year old and I’ll show him something a few times, you know, say this is a spoon, and he’ll suddenly be able to identify a spoon even if it’s 2D on a piece of paper or if it’s upside down. And he’s developed this model of spoon and it really is kind of the same thing. So we’re just developing what is the sort of model and character of a carry or of a periapical lesion or of a filling or a crown or whatever the case may be?
And so when you’re talking about 30 neural networks, you’re talking about essentially 30 different ways or models of interpreting that, bringing them all together. Yeah. And then delivering the result.
Yeah. And we have all types of neural networks that don’t even look at pathologic data. We have neural networks that determine, is this an OPG or is it a right wing or is it a periapical x ray? I mean, it’s doing all types of meta data analysis to know we do things around rotation and orientation. We’re able to understand tooth numbering. We’re able to actually segment out. This is a very useful tool for practitioners. We’re able to segment out anatomical structures so we can tell the dentine from the enamel, from the cementum, from the root, and we’re able to actually overlay that data on top of radiographs. So when you’re explaining to a patient, Hey, you really want to address this decay before it touches the nerve, because if it touches the nerve, then we’re going to have to engage in a root canal is going to be much more costly and much more expensive. Then you are communicating to a patient in a way that’s very clear, is very visual, and we actually dramatically increase case acceptance as a function of that because the patient now really understands what’s going on and it elevates the level of trust as well.
I can imagine that that side of it, the patient communication side of it, is probably even a bigger driver than I think it’s 5050.
I mean, they’re everyone focuses on the pathologic detection side. But I agree with you in that the patient communication side is just as important.
Well, you know.
So many of us suffer with that. And it’s not it’s not about dentists generally, not that great at communication, to tell you the truth. But it’s a difficult thing to see. I mean, it takes years. You know, you really don’t know what is decay and what isn’t decay when when when you’re a young dentist and then you try and show a patient and say, hey, see this little grey area here? Yeah. And most of the time they can’t see that exactly. Whereas if it was there in pink or something.
Would change the.
That’s the colour that we use is pink.
Every, every detection is a different colour, but we do different shades of pink depending on if it’s enamel, only if it’s actually encroaching into the dentine and stuff like that. But you’re right. I mean when you show a patient a grey smudge on, I think that’s already confusing. They’re not really going to understand what it is that you’re talking about. So this plays a huge role in patient communication case acceptance, and it’s one of the reasons why dentists, you know, sort of love using it.
Were you surprised when you moved into the sort of medical area about how many hoops you’ve got to jump through on the whole regulatory side? Because I’m in I’m in dental supply as well, and I feel like it’s a weird double edged sword. It’s kind of the the worst thing and the best thing about being in this area because it’s a nightmare to get the regulatory regulatory right. But once you’ve got there, then you know, there’s a real barrier to entry for anyone else. Yeah. However, however long it took you or however many millions you spent getting all of those FDA and I guess, you know.
I guess we’re also the.
Only cleared product really in the market. And that was.
What about Japan? Did you manage Japan? Because that’s a nightmare for everything.
Was a very difficult and very costly endeavour. And and you know, even the philosophy behind something like Emdr is very different than FDA or Pmda in Japan in that you’re looking.
You know, did you.
Know you were getting yourself into that?
No, I don’t know anything. So that’s kind of my that’s my strength, is that I’m just naive. If I if, you know, the pain you’re going to be walking into, you might not do it at all. So, you know, I historically have entered categories that I am not an expert in and naivety helps. I knew it would be a heavy lift and painful. Of course, I didn’t know quite in which ways or how. And when. You’re dealing with so many regulatory bodies concurrently and you’re still, you know, we’re not pre-revenue anymore, but we were pre-revenue for a while. You know, that’s a that’s a it’s a stressful endeavour because you’re spending a lot of money, you’re burning a lot of capital, and it’s kind of an all or nothing thing. Like if you don’t get through, you need to start from scratch, right?
Well, it’s bouncing around in my head and we’re probably a million miles away from there right now. But sort of early detection for things like cancer or perhaps rare cancers. And and I guess you need the data to feed the model, as you were describing then. But, you know, even if I go back to my own personal story, I’ve got an l5-s1 disc tear to the right, and it took three scans for for somebody to diagnose me, right? And actually, the third guy could look at my first scan and he saw it like, stuck out, stuck out to him like. Like a so, so, so how how often does this happen? In the real world. Right. Let’s talk outside of dentistry now. Right? All the time. And then and then you think just how powerful this could this be? That if. Well, early detection, cancer. Right. I mean, that would be insane.
Yeah. And that’s happening very rapidly throughout medicine. I mean, this is going to be applied holistically across the board in a relatively short period of time, and it’s going to be hopefully resulting in much better outcomes. One of the nice features of whether you’re detecting cancer or caries is in our case, we’re proving that it’s almost we’re we’re servicing almost 40% more stuff that’s getting missed. It’s a lot of stuff, right? But a bunch of it’s going to be, you know, say like in Interproximal Carry that is just an early watch area. There’s nothing that you necessarily want to do about it. You want to be aware of it and you want to take preventative steps. And that’s a nice conversation to have with the patient as well. They don’t need to do anything differently other than engage in better oral care. And if you’re able to capture things, catch things early, especially cancer, then you’re able to have much more elevated outcomes. So when you catch something is critical. But yeah, I mean, that’s a great example that you bring up.
How far do you think we are away from that?
I think it already I think the technology already exists and is sort of probably, if not already FDA cleared is in a clinical trial phase. I think there’s about 200 FDA cleared devices that are able to look at radiography and detect things. You know, mammography is another area that’s been very popular and has seen success. So it’s all happening. It will all happen. This is kind of not a question of can the machine do this? Well, that’s not an outstanding question anymore. We know that it can. So, you know, people are getting after it.
I saw one of your main investors is David Saxe, who I’m a massive fan of from the All in podcast. Yeah. When you I mean just on the general thing from from the sort of founder sort of angle when you’re pitching to these investors to start with, I guess you’ve got to pitch like you’re going to you’re going to take over the world. You’re selling a massive dream, right? Was the dream that you sold dental radiography I or was it I in dentistry and is that the plan?
It’s really about elevating the standard of care across the entire field. So at our core, what we do in the first step is to get this plumbing and make it a utility in every dental practice globally. But that is really just a stepping stone to a huge amount of other applications. First of all, there’s a lot of modalities. So you have the 3D realm and you have CSF. And one thing we haven’t really talked about is that we’re actually able to correlate the data that we identify in the imaging system with the data and the practice management system and identify all, you know, like all the characteristics of the entire patient population. Do a comprehensive chart review, show all the undiagnosed opportunity, show how that leads into various specialities, whether that’s endo or Auth. I mean, there’s a massive amount of interesting work to do here.
With I mean, it could.
Even go to treatment planning rather than just diagnosis. Right?
Absolutely. Yeah. And that’s one of the things that we do with our practice intelligence platform is we actually will colour code the schedule every day and we’ll highlight areas where there’s an action to be taken, both driven by AI and not. And we’ll list out in various funnels the various appropriate potential treatment opportunities for a particular set of pathology that are present.
Did I hear that right? Could could the software retrospectively go into the practice management software, hunt around, find all the x rays? And then pick stuff out.
Pick all these patients.
Exactly. Previous patients, thousands of patients.
We go back by default, 18 months in time. But we can go back. In some cases. We’ve done, you know, 5 or 10 years and we basically highlight the characteristics of that patient population. So that’s where we’re actually starting to cut the stake for the practice, right? And that tool set really services management, IT services, front office hygiene and GPS. So you’re not just providing real time detections, you’re actually taking that analysis, you’re correlating it with this other data set about the characteristics of the patient and the other work they’re engaged in, what’s planned, what’s been scheduled, the notes, all that stuff. And you’re holistically bringing it all together.
Tell me about the sort of the offering to the dentist. Is it is it the monthly subscription? Is that how they pay for it? Does it integrate with the software? There? Practice management software? How easy is it to sort of onboard? You know, a lot of dentists are tech savvy, but a lot aren’t, right?
So it’s a monthly subscription, we think quite affordable relative to what we’re surfacing. In other words, typically, how much is it?
How much is it?
So I’ll give you the US dollars. It’s 299 US for second opinion per month and 595 for practice intelligence. However, we are surfacing typically thousands of dollars per week in incremental restorative opportunity, for example. So really you can pay for this whole thing, you know, within for the year, within 1 to 3 weeks. I mean, it is a is a big sort of production oriented ROI there. And we intentionally wanted to price it very accessibly because we do believe that this belongs, you know, in all dental practices globally.
And is that is that a per practice stroke per clinic fee or per practitioner or how.
Does that work? That’s per clinic, per practice.
So it doesn’t matter if there’s 20 dentists working in there or five dentists working in there. It’s a flat fee basis.
Typically we’re not really, but there is a limit at which if you’re like, you know, 30 operators, one of these kind of outlier practices, you might pay for a few licenses, but we’re pretty generous with the image counts. So I would say up to five practitioners. And then, you know, you might kick on another license. But if you’re a DSO and you have 100 locations, that’s 100, you know, typically 100 license.
Of course, of.
Course. And then what’s the distribution model? Are you I mean, have you got your own sales team in the US and you’re working with distributors abroad or how are you doing?
Good question. So basically, we have a variety of ways to access this technology. We have our standalone tool for second opinion, which integrates with pretty much every major imaging system and PMS out there. So you can just kind of subscribe and use this tool. Increasingly, we do have partnerships with a range of PMS and imaging partners out in the market and they’re engaging in more deep integrations where you’re able to actually access these capabilities just natively within your existing interface. So just one example of which there are many, you know, would be like Planmeca. We did a big announcement with them and they are integrating these capabilities into Amex’s you natively and directly so you don’t even have to go to another interface. You could easily turn it on. And from a distribution perspective, we have our own sales team both domestically and globally, but we also work with distributors in the UK. We work with Dental Directory, for example, and you know, we’re working with the various sort of channel partners to get the thing distributed.
I’m really curious about your backstory, right? I mean, I’m just sat here in awe blown away by what you’ve done and what you’ve achieved. But tell me a little bit about your backstory. You know, where you grew up, what sort of school you went to, what kind of kid you were, and how you managed to navigate to to where you are today. Were you were you some computer programmer type kid? Just just talk me through. Talk me through your upbringing.
Yeah. I’m actually very fortunate in that, like, I knew, I think from a fairly young age what my interests were and how I could apply them. So I was probably 13 or so and I was like, I think I’m going to run venture backed, you know, tech tech companies. And it was very much a programmer, sort of obsessed with all things technology and specifically programming. And that’s ultimately why what I ended up studying and I got my bachelor’s and master’s at Carnegie Mellon and it was just very clear on what I wanted to do. So I really feel like I’ve been at this for since I’ve been like 14. Wow. I actually started a company in in high school, which is kind of a development oriented interactive agency and, you know, sold that. And it’s kind of a. A typical story for people that are often sort of do what I do, which is you kind of identify young and start young and have the entrepreneurial bug. So for me, it was never really a question. And yeah, still, I still love it. I’m still fascinated by it. I don’t actually, you know, have hands on keyboard sort of programming day to day. I moved away from that quite a while ago.
When you were about.
No, no, no.
I did. I did have one job in my life, um, working for a hedge fund. I worked for a hedge fund called Bridgewater Associates out of college, uh, out in Connecticut. And I was a programmer there working on trading systems and whatnot. Did that for like a little over a year. And, and then I went and started a company and then I’ve been starting companies since.
Okay, you don’t need to do this. What drives you now? I mean, surely the previous, you know, the gum gum success story, you could be sitting on a beach.
Yeah, it’s a good point. I mean, I’ve actually been having this conversation. I don’t know that I will do it again in quite this way because it’s so much work and it’s it’s a pretty painful process. It’s never not hard to start a new company, especially when the expectation is to frankly, you know, quickly grow to $1 billion valuation and more. I mean, that is the hope and the expectation on the part of the financial backers and also on my part as well. So it’s a very intense sort of all consuming process and it’s probably why a lot of people in their 40s, you know, that have had success in their past, stop starting companies and go more so to the investment side or or do other types of work just because it is so consuming. But in particular with Pearl, I really felt that I was sitting in this very unique position where we had a lot of proficiency with this technology. It was very clear that this technology was going to have a massive impact on humanity and on medicine. And I just wanted to play a role in in sort of applying this technology to more impactful and meaningful effect. So I’m very proud of what we did at Gumgum. But a lot of it was selling advertisements, you know, and I became known as like a media guy, and I never felt like a media guy ever. And at some point I was like, okay, I think I just need to do something different before, you know, my career becomes a different kind of thing. And I did go to the board and I pitched him on the idea of spinning out of me going and running it. And, you know, that was surprising. And there was resistance to that concept because it was unexpected. But I think we created a scenario whereby it was it was very win win to go and do something that I thought would impact humanity more fundamentally.
So listen, going forward, then, what are you looking at in sort of a five year horizon? I mean, do you feel like your job will be done by then or how long will it take?
I don’t think the.
Job will ever be done. I mean, we have probably a five year roadmap that we’re engaged in. There’s a lot of obvious work to do, and then there’s less obvious work to do. But it all effectively revolves around how can we apply different forms of cutting edge AI and technology generally to help elevate the standard of care in dentistry and to help unburden dental practitioners? You know, dentists in particular have very, very stressful lives. My father I witnessed this, right. So you go to dental school, you become hopefully a good at practising dentistry, but suddenly you need to find real estate and you need to run a business and you need to do hiring and firing and you need to do back office and accounting. And it’s just a lot. Yeah. And, and.
You’re working, you’re working in it, not on it. That’s, that’s really a.
In it. Not on it exactly. And that’s just a lot. So I think that technology can be brought to bear increasingly to just help unburden practitioners and also elevate the standard of care for for practitioners, for for the patients.
What would you say is your superpower? I mean, how many how many people did you end up having in Gumgum?
Oh, uh, gosh, probably when I left around 650 or 700 or so.
So so so as a as a CEO, I guess you’re spending a bunch of time raising cash, then you’re spending a bunch of time selling the vision to everybody, Right. Including internally as well as externally. Right. What is it like? I mean, you’re you’re clearly a technology, you know. King as well. What is it about you that you know from your from your the way that you’ve worked, that you think is being your superpower in getting these companies off the ground, billion dollar valuation, so forth?
I think that is my I like building companies. I like taking ideas and making them a reality. I think when companies become about more so operational driving, operational efficiency, I get a little bit less interested. That’s definitely the stage that gum was at two. It’s just.
I don’t think I’m the best person at that, first of all. And I also am not the most interested in that. I’m more interested in 0 to 1 than 1 to 5, although it’s much more fun to be in the 1 to 5 mode because things tend to be sort of working and humming in in relation to raising money. For Pearl, it was a really different experience than gum. So gum gum we raised like, I don’t know, $130 million or something. When I raised my first, you know, 500 K, I was a nobody and had to pitch everybody and kind of sell the vision and and had to do a lot of that. And because I have a reputation and we built a successful company at Gum gum and have relationships raising the money for Pearl was one conversation with one person, literally did not speak to anyone else about it. So Dave David Sachs came to my office to just check in. Yeah, it was. It happened to be around the time where I decided to do this spin out and I was like, you know, Hey, I’m thinking about doing this thing and spinning out this company. What do you think? And he’s like, I think you should do it and I want to back it. And let’s just agree right now, like, name a price. And we shook hands and he did it.
How much did he give you?
How much did you give? You was right at the.
Easier when you have relationships and some success.
Under your belt. How much was that for? How much?
How much did he give you in that handshake?
Before before you had.
Anything? No, because we had.
Incubated the idea. So we had some technology and some proof of concept, but we didn’t have much, to be honest. And frankly, I was probably overvalued relative to what we had pretty significantly. But he had faith. And his perspective is, you know, back good people back, good ideas, backed big markets and good things will happen, hopefully. And subsequent to that, we’ve brought on a number of other premier investors and it’s actually been incredibly easy and they’ve actually come to us. So that is not the typical story. I’m definitely very sympathetic to the other side of that reality, which I spent many years experiencing firsthand, but things have become markedly easier in that regard. It also helps that we’re doing this thing that everyone’s excited about in health care. So it’s a very bright spot within tech right now.
Yeah, as in as in from the from the investment perspective, people are still willing to invest in AI and health, whereas they’re not willing to invest in a whole lot of other tech, Right? Is that what you mean?
It’s been a difficult it’s been a weird go at the whole from a market perspective since we launched Pearl because we had Covid and that was, you know, crazy. And then we had a bit of a financial meltdown. And, you know, then getting to the hopefully what is now a recovery. It’s been a very odd set of realities on a macro level.
So yeah, let’s get to the darker part of the podcast. We like talking about mistakes on this podcast, and a lot of times with dentists we talk about clinical errors that they’ve made. When I say mistakes, what comes to mind?
I mean, as it relates to Pearl in particular, I just think that the jury is still somewhat out on it. In other words, we received our FDA clearance in March of 2022. We’ve been commercialising to great effect since then and we’re growing very, very rapidly now. But I think that we have taken a certain strategic approach to the market that is ultimately going to determine how successful we are, you know, relative to anyone else engaging in the space. And I believe in our strategy and I stand by it. But I might find that there were some real errors there because you can’t do everything at one time. You have to kind of pick your lanes. You have to pick your distribution channels and you have to pick your partners and you have to pick the technology that you’re going to focus on, the problems you’re going to solve. There’s a lot of decisions that go into that with very imperfect information. I’m sure we didn’t do it perfectly, but I think that’s really going to sort of the mistakes that we made will rear their head over the coming years, and it’ll be more clear than right now. Right now, it’s such a greenfields opportunity. There’s so much demand, it’s so new, the market is so ripe that it’s hard to really tell. I can get into certain approaches that we took technologically, which ended up being dead ends, but that’s.
The price of progress. Yeah.
Yeah. So we took certain approaches technologically from a from a machine learning and training perspective that, you know, we’re, we’re wrong that we had to backtrack on and do better. I think that the way that we approached regulatory was, in retrospect, incredibly bold with the FDA. For example, we went after ten clearances at one time. Wow, nobody’s ever done that before. It ended up working out, but I think it was also a big risk to have to have done it that way. So I don’t want to say we got lucky, but it was definitely a risk and that was a function of not really knowing on some level what we were doing, even though we were sort of advised, I don’t know. I don’t want to I don’t want to give you a I don’t want to shy away from the answer.
I just had a mistake. It Yeah.
If you have ever had one of these as an entrepreneur. Right. And look, I’m a very small business owner, right? But I’ve had numerous falling down moments, right? Oh, shit. I’m in. I’m in the deepest, darkest hole. Right. How the hell am I going to get out of this, right? Emotionally, emotionally, you’re in that space where you don’t know whether to cry, laugh, break down, whatever that is. If you ever had those moments and what were they and how did you get out of the hole?
Oh, so many times. I mean, so often. So. So I’ve had those moments a lot, which is why it’s such a difficult and miserable endeavour to start a company like this. So with Pearl, it’s an easy one. I mean, there’s been plenty of those moments, but, you know, the most obvious one would just be, okay, now we’ve spent, you know, X amount of millions or tens of millions of dollars, and if we don’t get this regulatory clearance, we literally cannot operate. And I will have failed. I would have lost all the investor capital. I would have been a fraud and all those feelings, Right. And then that same thing comes to bear when you’re entering new countries and big partnerships or, you know, there’s just, you know, while we’re doing well at the moment and everyone’s very excited, like it does not feel like we’ve punched through into being a big company by any stretch. So that’s a very familiar feeling. It was a very familiar feeling at Gumgum. It’s all the same exact feeling. Everything that you mentioned, Yeah, it just coming in different in different forms. But yeah.
You’ve got a two year old kid, was that right? Did I hear that right? Yeah, yeah. Just talk to me about work life balance and what a day in the life of fear is like and how do you balance your duty and your role as a father and making that time and space. Maybe you’re one of these guys who manages it incredibly well and what a typical day looks like and how you manage your time.
Yeah, well, my partner, she’s a she’s a filmmaker, so she’s a writer director. She actually just made a movie. So she’s been in the editing room for the past three months. Every day. Yeah. Um, I actually so in during Covid, we became a remote company and that was actually pretty functional for us. There are some drawbacks of it, but while the majority of the executive team is in Los Angeles, we now have people kind of everywhere and we operate very remotely. So I do primarily work from home, um, which has actually been wonderful because that means I’m able to see my kid a lot in a way that I never would have been able to before if I was working. Pearl Like hours from an office, that would be difficult. Um, so that’s been a really nice feature of that reality. I’m able to pop up and, you know, spend 5 or 10 minutes and that makes a big difference. I would not say there’s a, it’s a I would not say there’s a very good. To work life. It’s kind of always the work is always on. I’ve gotten good at separating the two, at not letting you know the the current feeling about about the company sort of overshadow everything else, which is a skill you have to have, I think if you’re an entrepreneur, but it also makes it so that you know the two year old, it’s such a fun time and they’re so joyful and they’re so cute and it’s just actually like a great escape. So I feel pretty good about it and I feel very fortunate to be to be remote in that regard.
So I’ve got another question related to that, and it relates back to me right when Covid kicked off and we were forced to work at home, it was such a beautiful time because I was forced to spend that time and be present with kids, with my wife. And even till today, I will say for me personally, it was a bit of a blessing because I connected with with my kids and my wife in a way that I wouldn’t have done. I definitely wouldn’t have done during that time, right? Yeah. But then that became the norm. So we went back to work and perhaps started working from home and we became a remote company, right? And then what happened is the divide between work Prav husband, Prav and Dad Prav almost amalgamated into one. So this was very difficult for me to mentally shift between being work Prav and Dad Prav with within like 10s right? And then husband Prav and then walk into this home office where I’m sat now and then become the work guy, right? Mentally, I struggled with that to a point where in January I just had to get myself an office only 15 minutes from home. Right? But to make that mental shift of I’m going to work and then then the guy who’s going to walk through the door at the end of the workday is your husband and your dad. And and that that for me personally, I needed to make that shift because I felt like I was blurring the line between that and I wasn’t being challenged. Wasn’t being present. Do you do you feel that you just mentioned you go and like steal like ten minutes of joy here.
And provide office.
Space options for all of our employees for that reason? So if they want to go work out of the office, we provide them the ability to do that. For me, having worked in an office for so many years and doing the commuting in LA, I happen I’m fortunate enough to have a guest house that’s set far away from the main house that nobody ever goes to. So I.
Very quiet and I’m uninterrupted. For me, it’s more so about state of mind like, yeah, it’s not about the amount of time. It’s really about being in a good frame of mind, being present. And that’s where it can be challenging when you’re preoccupied with with other things.
What’s the competitive environment like? There must be there must be competitors. Who are they?
There are two competitors with FDA clearance in the United States. They don’t have clearances globally. So, you know, we’re competing on one way domestically and in a different manner globally. Certain other countries like basically have very like if you look at Australia and there’s not much not much else out there other than Pearl. And if you look at the UK, it’s more so regional sort of efforts often coming out of university with some subset of capabilities. Nobody’s really brought the kind of capital and firepower that we’ve brought to bear on this challenge. But you know, there are, you know, an array of competitors out there and, you know, there also an array of countries that have different sort of areas of focus. So I do think that there’ll be some confluence of companies that do certain things well that will help push those capabilities into the market. It won’t just be Pearl, you know, that’s not what we want. But I would say that it would be really hard for any new entity to enter the market now relative to where it’s at. It would just take years by default and a lot of money. So even if you’re a very large company, very committed to this, you probably have to buy something and and approach it that way versus build.
When you say people focusing on different parts of the market, do you mean some people focussed on dentists, some people focussed on DSOs? What?
Well, more so. Like some people might focus on applications for payers, right? For insurance companies, or they might focus specifically on C.F. or CT and not so much on 2D, or you have companies like Dental monitoring who’s really ortho focussed and not really focussed on radiographic anything, but more so on their own forms of AI and stuff like that. I mean there’s a lot of or you know, you have smile design which, which is employing AI to great effect. You have all types of laboratory applications that are, you know, applying AI for, you know, designing. Aspirations and stuff like that. There’s lots of stuff.
So before we looked at some of your darkest kind of days, when you look back on your career, what are the sort of the highlight days? What comes to mind when I say that? The high points. Was it like selling these companies or.
The high points for me are always pretty much the same. I remember the day where I realised that gum Gum had a thing that the market wanted enough that I could repeatedly provide to them, and I was like, okay, now we can just do this. And there’ll be a lot of challenges and growing pains, but now we know what to do. You’re not kind of meandering in the forest trying to figure it all out, right? And I would say that that’s a common experience across my companies and that I typically tend to do things that were historically never done before and unproven, and it was unclear if they were possible. We had a similar moment, I would say, at Pearl not even that long ago, where I was like, okay, this feels is like a more comfortable place to be because we know what we’re doing and we know that the market wants it. We know it’s good enough, all of that. So for me as an entrepreneur, that’s always the best moment because it’s like a real release that is often pent up for years. But you know, a lot of highlights along the way. Of course.
You said before.
When you’re looking at when you’re looking at all or nothing with the regulatory and you’re thinking about the investors money and that weighs on you, does it weigh on you when you have 700 employees or do you not? Are you not wired that way? You know, like are you thinking. Are you thinking that, you know, all of these people’s lives are dependent on whether we make it or not? Is that not in your thinking?
Is the question Is the fact that we’ve taken on a lot of capital and how will this responsibility kind of weighing on me all the time or not?
Is that capital?
Yeah, but but also the number of people, people that you have.
Yeah, it’s a lot of responsibility. And not only that, but it’s really like it’s responsibility to people who are giving like the entirety of themselves to the effort. So like, you really don’t want to let them down because they’re giving you so much, right? And you feel a lot more responsibility because of that. This is not your typical 9 to 5 work. This is like we’re all figuring it out together. We’re pulling weekends or pulling nights. And and yeah, I mean, like I said before, I think you have to get pretty good at sort of managing that level of responsibility and also separating it out and on some level realising that like you’re doing everything you can, you’re doing the best job you can do, and sometimes you just need to like let that be enough rather than drive yourself insane. But yeah, it’s all it’s all a big deal.
But you must have had moments where you’ve had some employee who’s really, really pulled weekends for you, laid their lives down for you, and then some. For some reason you have to let that person go.
Yeah, many times.
And that’s how do you deal with I think that’s like the hardest thing in all of business, right? Because it’s mean. It’s even hard to let people go who are terrible. But when someone’s been really good and really tried their best for you and laid their lives on the line for you to let that person go.
It’s the hardest thing to do. I feel that I have done that many times.
Yeah. I mean, over, you know, a thousand plus employees and you have a lot of people that are very dedicated and their heart’s in the right place, but they’re just the wrong fit for your organisation for any number of reasons. Either they’re a skill set thing or it’s a talent thing or it’s a culture thing, right? And the best thing that you can do, in my opinion, is be honest, be kind, be generous with those people and try to be helpful to them. But you have to protect the enterprise as a whole. And you have to make, in my opinion, very swift decisions as it relates to the to the personnel at the organisation and this is why I don’t like the family analogy for companies, because you can’t fire your family. You know, I’ve tried but.
But you know, a.
Professional sports team, which is a good analogy that you know Reed Hastings uses a Netflix I’m kind of stealing that, but I think it’s the perfect one because we’re recruiting top athletes and we’re expecting a lot from them and they get released if they’re not contributing in the way that we need them to. And that’s a much better analogy. That’s much more accurate, I think.
And who who’s inspired you in business or who are who are the people you look up to in business?
You know, I find myself thinking, well, there’s a lot of people that I sort of more current that I know personally that I look up to a lot. But I find myself very cliche, I’m sure thinking of Steve Jobs a lot. I mean, what a phenomenal visionary. I know that he was very hard on people and could be, you know, a real. Hole and all that. But I think his level of of vision and commitment, like I understand where he’s coming from when he’s flying off the handle because he cares so deeply. Now, that might be the wrong human oriented approach, but I kind of understand and I’m sympathetic to like what he’s going through internally because he just wants to bring this thing to the world in a very, um, in an elevated manner. So, you know, I think there’s, there’s obviously many, but that’s one amazing.
I think we get to the final questions now.
Right Let’s get to the final questions.
Um, so we usually close off with, with the same final questions or fear and so, so my question is this a fear you’ve, you’ve conquered everything that you can in dentistry and I and and the business is financially done everything and you’ve achieved everything in life and you come to that point where it’s your last day on the planet and and your, your your little one is, is next to you and you’re surrounded by your loved ones and you have to leave three pieces of wisdom. What would they be?
First of all, how dare you for asking this hard question. I was telling you, Payman. I hope they had it sooner, but I’ll do my best.
So I guess three pieces of advice. One would be. One for humanity generally, which I think would be well served to remember, which is it’s extremely miraculous that we’re alive and exist at all today. Like floating on this rock. And an infinite number of weird circumstances have led to the fact that life is possible at all. But the fact that we’re also living at a time of general peace and prosperity is just insanely fortunate and unlikely. And I think that to be driven by gratitude and love and light of that reality day to day is is well founded. And an important thing to keep in mind because we get very myopically focussed on, you know, the next thing or making more money or, you know, achieving and all this material stuff. But just the fact that we’re here at all is a real gift.
So do practice gratitude like in a in an organised way. Do you do it every day or something?
Oh, man. You know, one of the sacrifices that you make when you build a company like Pearl, because I got to that point at Gumgum, I had all this like time in a way that I got back to do the things that I wanted to do. That’s the nice thing about having a larger company is you can kind of focus where you want to focus. You have an amazing team elsewhere and you can get back to being a human being with a life and in particular interests. And you get into yoga and meditation and just like self-healing, you know, and all this stuff. And the truth is, all that goes out the window largely when you start a new enterprise because it’s gruelling and you’re just like, it’s painful, right? And there’s a lot of suffering. So I think that that’s the real cost that people just keep in mind if they want to start a business, that’s the sacrifice that you end up making. That is a very significant one, and nobody is really spared from that. I don’t think you could be somewhat better or worse in managing it, but you’re not really getting out of it. And you know, when you care a lot about something, I don’t think people mind working hard. But if you’re like, you work really hard and the thing is not realised, there’s just a certain type of pain associated with that that I think is challenging uniquely. So I guess the second one, I don’t know. I mean, I guess I’m getting back to it on some level, but I’ve probably been kind of bad at actively doing that. I’m playing more tennis. You know, I have a wood shop, so I do a lot of woodworking. I could always tell about my like the state of my mental health is a function of how often I’m in the wood shop.
That’s always a good time.
So you know a lot more now than it was, you know, a year and a half ago, say And another piece of advice, I guess, would be to find things to be genuinely interested in and vigorously pursue them. I think that I know a lot of smart, interesting people that are just not interested. And I think that’s a real struggle. And I feel for those people because I think there’s something they don’t want that it’s just that they can’t quite find it. But to the extent that you can, then I think that’s an important part of life. And then also just to be courageous in your decisions, because I think that fortune favours folks who are courageous and there’s obviously inherent risk in that. And you want to be smart alongside being courageous. But my guess is on a deathbed, a lot of regret will come as a function of not having been courageous.
Very true. Very true.
And and so how would you like to be remembered or fear was and finish the sentence.
My wife always says, you know. You’re not necessarily like super nice, but you’re very. But you’re. But you’re super good.
Um, and I think that’s accurate.
I’m not the I’m not mean at all, but I’m just very blunt. Um, and, but I’m, but I do feel like I’m good. Like, I want the right good things for, for my people and humanity. So maybe that.
Blunt, good guy.
Yeah, the blunt.
Very cool. And then. Hey, do you want to finish with yours?
Yeah. We’ve got a fantasy dinner party.
Yeah, right. This is a tough one, too, because giving the hard question. So three.
People you want to spend.
Time with, dead or alive.
And I assume that these people could come if they come from different eras, the same language and conversation and all of that.
Yeah. Yeah. I mean, my.
Mind immediately, just a science nerd in me goes to like Alan Turing and Lady Lovelace and Richard Feynman and Isaac Newton and people like that. But. I guess.
It depends on what you’re optimising for.
So that’s a certain kind of optimisation.
If you’re optimising.
For something that’s historically interesting, maybe like Jesus or Moses and like Julius Caesar and like, you know, like Washington or something like that. But I think if you’re optimising for just like something that is highly entertaining, maybe. Fran Lebowitz. Christopher Hitchens. Like Einstein, I imagine. I feel like he’s a very.
It would be fun.
It’s a question.
But an interesting one. Yeah.
It’s been a massive pleasure to have you on. It really has been great.
Thank you for your time.
And I’m feel pretty sure Perl is going to do very, very well. And I can see by the team in the UK are doing a great job. You know, they they’ve gotten that name out there and they’re they’re at all the right places. So it’s good. It’s good to see that too. Really, really massive. Pleasure to have you, buddy. Well done. Good job.
Thank you so much.
It was a real pleasure. It’s a unique podcast and appreciate the thoughtful questions.
Thanks a lot, man.
This is Dental Leaders, the podcast where you get to go one on one with emerging leaders in dentistry. Your hosts. Payman, Langroudi and Prav. Solanki.
Thanks for listening, guys. If you got this far, you must have listened to the whole thing. And just a huge thank you both from me and pay for actually sticking through and listening to what we’ve had to say and what our guest has had to say. Because I’m assuming you got some value out of it.
If you did get some value out of it, Think about subscribing and if you would share this with a friend who you think might get some value out of it, too. Thank you so, so, so much for listening. Thanks.
And don’t forget our six star rating.