AI in Dentistry: Insurance Verification with Vivek Kinra with Verrific

Welcome everybody to another episode of the Dental Boardroom podcast. It's Wes Reed, cpa, cfp, owner of practice, cfo, coming at you today with another episode in my AI in Dentistry series and this one is a little bit off the cuff. I just met this individual who I'll introduce you to in just a moment, literally minutes ago, and we had, we were starting to have a conversation about his company and what they do around, I believe, insurance verification. And I immediately could tell right off the bat that he knows a lot about AI and AI in dentistry and has some thoughts that I think will be really relevant for, for you dental listeners. So welcome to the show. It's Vivek. Vivek Kinra, is that right?

Yes, Vivek. Yeah, Vivekinra.

Yeah, Vivek. All right, Vivek, you know what I'm going to do because this was a little bit last minute, normally I do a whole prep and I look at your LinkedIn bio and all that stuff. I, I'm going to let you do the honors and just share about your background in dentistry and what you're doing right now as it relates to the business side of dentistry, computer automation, AI, et cetera.

A software engineer by degree, right? That was my career. My wife is a dentist, started a practice and in 2007 I really got immersed into it very, very deeply. In late 2018, I started a company that you may know called PPO Profits. PPO Profits was an insurance negotiation and billing and credentialing company. Became one of the bigger companies in this space and I owned that company for five years. In October 2023, it was acquired by Bancor Dental. And one of the reasons why I ended up selling the company is because I had also started this company that I currently own called Verific, which is more software driven insurance verification. It's a company called Verific. Insurance verification is like this one missing piece in the patient engagement cycle that nobody has been able to solve, you know, really for decades. People have been working on this problem for a long, long time. And so after my exit from PPO profits, you know, I've been full time the founder and CEO of the company and it's grown quite a bit as well. And in late 2021, we raised some money for an automation software and called Very Smart. So last year we released that software out of beta and you know, hundreds of clients are using it right now. And then over this year we have incorporated a lot of true, genuine AI into it. It's actually no secret, a Lot of the folks that actually know me well enough know that my development center is out in Poland and my dev center is in. In a remote city, that it's a physical center that I go there every year. But this year we also built out an AI lab in a city called Roslaw. And we've got a, you know, really, really talented team that's in AI, you know, very like, PhD level that you really know how to break AI into different tokens, et cetera. So really, really deep into the weeds of that. Right. And so we have, you know, when it comes to insurance verifications, you need to get very fetch data from many different sources, because there's no one source of truth and all the data that you need isn't in one place. Right. And so it's something called clearinghouse that you can pull it out, portals, you can make phone calls, et cetera. So, you know, verismart actually collates all that information together. Right. And so we have this new module called Portal AI, which is really just going to the different portals via, you know, a large language model to, you know, bring that data into our platform as well for. For the dental offices. So that's my background. Yeah, great.

That's a. That's a really good context to kick off here. You know, I think from our phone call, I was, I was. You triggered, I thought, a. A really good concept around AI in dentistry that I think this is true of all industries. To be honest, right now with AI, it seems like every company wants to somehow attach themselves to AI and put that on their marketing, put that on everything, to make it sound like they are really on this curve, this AI curve right now, and that their business is better, faster, more productive, better quality, whatever, because it has AI as a part of its processes and tech stack. So how much of AI as it's making its way into dentistry is genuine AI that's making a difference? And how much of it is just hype and claim to be using this technology, when in reality it's not really functioning there. It's not there yet. Are you looking to buy or sell a dental practice? If you're a seller, how do you find a strong list of potential buyers? There's no MLS or Zillow for dental practice sales. In such a fragmented market, many dentists selling their practice feel discouraged. That's why I built practiceOrbit.com practice orbit is modernizing how dental practices are bought and sold. It's a platform technology. Y' all know them, Zillow, Airbnb, Uber, and OpenTable, just to name a few. These technologies do two things. First, they help match buyers and sellers by centralizing the marketplace. Second, once the match between buyer and seller is made, these technologies help them navigate the sale to a successful close. Here are a few of the features inside a Practice Orbit users can determine a reasonable sale price for practice in just minutes. All they need is the most recent business tax return. Practice owners can use the directory to find important service providers like practice brokers, attorneys, accountants and bankers. Practice owners and their brokers can invite those service providers into the platform so all parties are coordinating the sale together. Practice owners will have a beautiful digital listing of their practice, like a home on Redfin. And importantly, the listing will be anonymous and personal information like your address, name and pictures are only released after an NDA is signed and approved by you or your broker. Practice buyers can create saved searches to be auto notified when new listings come on the platform that fit their desired specialty and location. And hey, dental practice brokers, you can list all your practices on the site for free. By doing so, you'll get access to Practice Orbit's large pool of buyers, all of whom have created profiles and saved searches to be auto notified when new listings come for sale. You can use the platform to introduce your listings to a targeted set of buyers. Additionally, and optionally, you can use the platform tools such as a built in NDA, a customizable LOI, and a workflow tool to navigate the sale from LOI to closing. In fact, you can manage all your listings on the system from listing the practice to NDA to loi, and finally to closing the deal. Create your free account today at www.practiceorbit.com.

Yeah, very little of it is really AI to be quite honest, right. You know, and you know, even with me, I've raised some money, right? And you know, when you raise money, you want to have a higher valuation, etc. I could easily just put AI stuff down when I was raising money, right? And get a stronger valuation, etc, right? These are all games that people can play, etc. I tend to be very, very precise in what I say and what I write, etc. Right? And so there are certain genuine AI applications, you know, phone calls, right? Like, you know, virtual reception is very, very, you know, all of that stuff is AI has to be right. Because if you think about what is determinism versus non determinism, I'm going to go a little technical in here, right? A lot of the software that you write has to be deterministic, which means that's the same exact thing every single time you run that piece of code. Meaning if you want a text message to go to a patient three days before the appointment with a certain text, there is no variance you want in that you want that text to go. And if you say, hey, please press text back C to confirm, you don't want variances of that to go through, okay? And it's, you know, so those are applications where you know, it's deterministic, does the same exact thing if you want a new patient to book online, right? There are literally there. There's a company I was talking to saying, you know, and then we've got this like AI online scheduling platform. Now I can dive deep, right? Because the moment I dive deep, the detail breaks everything, right? Because there is no good thing for people to say that there is legitimate AI that helps them schedule online. That doesn't make any sense, right? It's truly nonsensical because you have a few open operatories, you select which days and times that you want to be able to have patient schedule online. You can't have that modification done. And here's the thing, it's easy enough to write the code to do it properly. Why would you actually do it via AI that is non deterministic, which will indeed make certain mistakes? I mean, there's a, there's a. Try it. You know, you can ask the same exact question twice to ChatGPT, it'll give you a different answer, right? That's non determinism, right? Because it's. And there was a very, very classic case that circulated where if you asked, you know, chatgpt, how many Rs are in the word strawberry, right? And it was saying two. Why? Because you know, when you break it down, right? Like there's, it's applying context as well, right? When people ask how many Rs are in the word strawberry, there are actually context around it, meaning they're typically talking about the Barry Word, right? And so it's saying, well, there's two, right? In reality, that's not. I mean, there's obviously three Rs and there word strawberry, etc. Right? Now when it comes to a virtual receptionist, you don't want determinism because in that case it's just a glorified virtual IBR system, right? You want to have natural language processing, right? That's the whole, that's the hallmark of a large language model that you can apply tone to it, you can apply mannerisms, can give it personality, you can give it voice and It'll basically speak during that. You can give it accents, all that stuff, right, to make it as natural sounding as possible. So that is a very, very good application of true AI. Same thing with where X ray, where you're basically reading some X rays, et cetera. Because most of that done is via machine learning. Just feed it tens of thousands or hundreds of thousands or millions of X rays, et cetera, so that the software starts understanding that.

And it's really like large language models, machine learning, neural networks. Those are some of the ways neural networks get even deeper because you have to be able to make a decision. And a very, very good example that I sometimes give, again, maybe not very relevant to dentistry over here, but I want people to understand what is true AI. You can feed tens of thousands of images of a horse to a machine learning algorithm, right? And then you show it a picture of a horse and it can identify that it's a horse. And the way it does it, it just matches the contours, et cetera, of the neck, etc. And forms patterns to say, yep, this matches the horse pattern basically internally. Right, but that's not how a kid learns, right? You can basically show a picture of a horse to a kid and then you can actually draw a picture of a horse to a kid and you ask the kid, what is this? And the kid will say, horsey, right? That is because it hasn't seen enough pictures of horses, but it has figured out what a horse looks like, right? That goes more into the neural networks and everything, Right? So some applications are really good, but when I hear things like patient engagement platforms, the recalls and reminders and treatment presentations, etc. It's all AR driven, it's buzz, you get a higher valuation, you get some sexiness around, you know, branding and marketing. Right? But it's not reality.

So on that note, is it safe to say that in a, in a process it has an AI element to the process? If somewhere in that journey of that process it's dipping into an LLM which is then producing an entirely new response or output from, for that thing. It's like a brain that had to think about it. And it wasn't just working off of very direct mapping and very specific deterministic coding that produced the result that was pre anticipated? Because that's just really good software. That's just really good software. It's coded well, everything maps well and it gives that exact same response for the same input every single time. And you want that in certain processes, but other processes, it needs to be not this natural language. It needs to be brand new, it needs to be very tailored or customized, which is going to require the LLM to do some thinking around that part of the process. Is that a safe way to explain the difference between an AI driven process is one that dips into the LLM at some point along the journey, produces a brand new result that may have never been that way before. And that's different from just a really good coded process that follows the same thing with the same response every time.

It's very, very good way. And I can kind of like really, really, you know, give a very, very good example because I live that reality almost on a daily basis. Right. Like we built out an AI lab, right? And we're basically getting the information from the portals and it needs to get transformed to a particular format that we've got.

Can you clarify what those portals are again?

Yeah, sorry, sorry. Yeah, yeah, I appreciate it. We're talking about insurance portals, right? Where the insurance policy details are, right? Every employer has different. And there's about 230 different insurance companies, right? Like Delta Dental, Blue Cross, MetLife, Aetna, Guardian, right. The data, the way they store is all very, very different, how much data they have, it's all very different, right? So if you want to standardize it, and I'll give a very, very. I'll give an example where if some dentist is, you know, you know, listening to it, et cetera, which is probably most of your clientele, et cetera, they'll get what I'm talking about. There's a very particular clause in a dental policy called missing tooth clause that you got to be care, right? And so you want to, you know, one of the first things that people want to ask when a new patient comes in is, does this policy have a missing tooth clause? Right. And, but the insurance is, when you go into the portal, it doesn't say missing tooth clause, yes or no. It'll say something cryptic like prior extractions are not covered or cogenitally missing teeth, blah blah, blah, whatever. And you just need to be experienced and trained enough when you're reading to say, yep, there's a missing tooth clause in this policy, right? So those are areas where when we fetch that data, it's good application for AI to say, okay, we can't possibly give it every single different permutation or combination, right. However, the AI is smart enough to just understand that language. It also has broad, depending on which model you use, right? It has broad context around it, right? Like you can literally prompt it you can give a prompt engineering to say, please pretend that you are at the level of Charles Blair, who writes the book on dental coding called Coding with Confidence. Right. And so now it's going with that context and when it reads a sentence, it now knows what it means. Right. And so it can then transform that data. So those are really, really, those are the ways that we use it now when pre good AI, which really became better this year, right. But we started building this application three years ago and at that time when we were getting the data, when we pulled the data from an insurance company and it was in whatever randomized format, we had to manually map it to a format that makes sense because that's what dental offices need. They don't want to use their brain to search for it and figure it out and process it. They want to have one format and it just automatically maps to that format. And we just wrote code takes a long time, but it is deterministic. Once we write that code, it works as intended. We then for fetching the data from portals, decided to use a large language model. And we tried many, many different models. We've got the OpenAI ones, Gemini and Deep Seq from China, all those different models. And it had to be trained. And it's still a challenge. It's an ongoing process because once in a while there's a very common concept in AI called hallucinations. When you put a large amount of data, it's just, you know, and so you have to like clean out all the hallucinations, etc. Now the idea is that, you know, when you're using, you're training your own model and it becomes smarter and smarter and smarter, right? So once you've gone through four insurance companies and you've trained it well, the fifth one gets easier. The six months it gets easier all the way to, you know, 230 ones, right. Versus just writing code for each 230 of those individually. But it's not as simple as, hey, we're just going to take this data and randomly transform it. It gets it wrong often.

And to be perfectly candid, I'm not convinced today we have put investment in our own AI lab. I am not today convinced it's the right way.

So the thought that I have on this is when you have a process and you control that process end to end and you know exactly every step in that process, you can build that deterministically, you don't necessarily need AI, but when there are variables in that process or stops along the way, that Run into something that has what feels like endless variables to it. So if you've got all these insurance portals and every one of those insurance portals is different and communicates differently, and then within each one of those, it may have a different way of communicating depending on who's on the other end. You can't really build an automated process that is deterministic out of that because there's just endless variables, variables. And that's where you need the LLM to be able to take in the information, think through it, and then respond in a way that is accurate as if it were a human. The other thought I have is, well, somebody mentioned to me AI is a little bit like an employee and you have to train it and it can take some time as employee. And so if I think about that, I have an employee and I need to take time to sit down and train that employee. And then employee gets better with every week and every month and every year. But at some point, that employee becomes incredibly proficient at that thing, maybe even better than me and doesn't need me. I see rolling out AI possibly in the same way.

I agree there that, you know, I am, you know, just cautiously optimistic there. Right. I don't, you know, I wouldn't say I'm cynical because I'm not. I follow truth and accuracy. Right. Whatever precisely takes me. If something is 7.3, I don't round up or round down. I'll just say 7.3. It's the engineer in me. Right. So where the current AI stands today, like, if you tell me that, hey, let's, you know, I am confident that this is the right mechanism and this trained AI agent is going to be better than an employee. I haven't seen it yet. Okay. Like, there are companies that we try ourselves, whether it is to make outbound calls to insurances to fetch the data. Right. That is a good use of AI. You know, you can, instead of having human beings to make phone calls to insurance companies to get the data, you could have an AI that can handle natural language processing to get that. Right. Similarly, you've got like, you know, inbound AI receptionists, right?

We've tried a few and I am yet to currently have a very pleasant experience in our own dental office. It just somewhat easy to trick it. And I'm not trying to trick it. It's just that the real world doesn't work very clearly. Right. Like, you don't just get a very happy use case that, yes, I'm a new patient, yes, I have met life. Yes, this is this, right? Like, you know, people will just simply say, oh man, I don't have, I know I have dental insurance. Can't you check with my Social Security number or something? Right. And these are all the different real world applications that happen that you know, very often. I'm finding that you know some of the big AI companies out there for inbound receptionist. I don't know if I'll put one in our office yet. And I've tried a couple, I've trained a few.

Do you ever feel lost in your practice finances? And how about your personal finances? Ever wonder where your hard earned money goes and why you're not seeing the first financial progress you've worked so hard for? If so, listen up. As a dental CPA and financial planner, I've created our company Practice CFO just for you. With almost two decades of experience, Practice CFO integrates great CPA services like accounting and tax with customized financial planning, both business and personal. Think of it this way. Our dental focused CPA services give us the the X rays we need to clinically plan your global finances and be effective financial planners. It's a highly customized approach we call the CFO model. Historically, having a CFO or chief financial officer was an advantage of large corporations only. But today Practice CFO gives that same advantage to private dental practice owners. As a client of Practice cfo, you'll be assigned find one of our dedicated CFO advisors. You'll receive insightful financial reports each month, have regular financial review meetings, reduce taxes, grow your wealth and feel more in charge of your numbers. Because we believe that in the competitive landscape of dental practice ownership today, you have to be an effective business owner, not just a capable clinician. Reach out to learn more at www.practicecfo.com.

So this is a really important point that I think about with AI. Think about it with my own business because there's now emerging AI based technologies for CPA financial planning investment firms like, like mine. In the life cycle of a new innovation around technology as AI is, you will have different types of users along that lifecycle. And there's a book on this, it was called like hyperscaling or something. But you have the original, you have the developers, but then you have the first users. The earliest of adopters. That's right. And the earliest of adopters are ones that are there. They're always curious, they're always wanting to look for the next level of incremental improvement and they're willing to almost trial and error or beta test these new technologies. And you know, in Software development. The concept of agile development is you build the skateboard and you sell it and then you add on a handlebar and you then sell a scooter and then you go to a moped and then a motorcycle, then a car and eventually you have, you know, like a F1 race car and a jet. And that agile development is along that way you have something that provides some utility. It just gets better and better and better. So as we think about the utility of AI in a dental office, are we at that point where it's still extremely early and therefore the people using it are the early adopters, or are we there yet to be able to start bringing on large scale number of users on it? Is it dangerous? Almost like Waymo. Like Waymo is winning the game because they have had enough experience and more data to collect and therefore they're well ahead of Tesla and the others because of that. But I know a year ago I wouldn't have got in a Waymo car. I'm not that level of an early adopter. I'd be ready to get in there now. So where are we in that analogy in dental?

Yeah, I mean, I think we're slightly past the early adopter, but I would say that a lot of the tech, the AI tech, and I'll go with the AI receptionist, to be quite honest, because those are the highest number of companies that I'm seeing all of a sudden that can kind of automate some of the mundane tasks. On that front, I would say we're more in the beta testing stage. The applications that are out there, I don't think they're ready for primetime yet, especially for a private office. Now it all kind of depends on your own tolerance as well. There are offices that just say, hey, I'm an AI assistant. And then, you know, there are offices that are like, yeah, I'm gonna have like a bunch of patients that might get turned off. Right. Just like they get turned off when they call and there's not a human being that picks up and you're like agent representative, et cetera. Right. But the cost benefit analysis for them. Right. And then there's a lot of private practices, more like boutique practices, that don't want to give the patient an experience that may frustrate them. Right. Because they just want to talk to a human because in their minds they have a nuanced question that the AI cannot answer. Right. Like people make these preconceived notions that I just need to talk to a human because there's no way you're going to be able to give me an answer about what I'm looking for. Right. So I, you know, there's obviously so much tech in dentistry that's like tested, bulletproof, right? You know, again, new marketing, online scheduling, recalls and reminders, treatment presentations, all that stuff is like locked in, right? Like where it's, it's, we don't have to take a second look at it. Again, it works really, really well. Like I said, I am probably fairly particular. Even though I'm an early adopter in most tech.

I'm not an early adopter towards production which may actually negatively impact the business. And so we have tried a couple of these AI receptionist companies ourselves. I am not yet ready to put it live, even though we need one. Right. So I would say more the beta testing stage. Now there are other companies like Pearl and Overjet that I think are very, very mature, right? But you know, I don't think that they're using a large language model. It's probably more machine learning, et cetera and I don't know the internals of it, right. Like, you know, I don't know if it's true, genuine deep machine learning AI or you know, what is it? Right? Because remember, right. Like there is and I'm not saying the Perlin Overjet, are they, right. Like they're mature companies and they're really, really big. There's like stuff that's gone through with like FDA clearance and everything, right? So I'm definitely not qualified to comment on them particularly. However, I do certainly see a lot of hype around, you know, claiming more AI. And I will not name companies here. But I can tell you that there are today companies that claim AI and it's just humans at the end of the day in dentistry. In fact, there's one company that claimed like oh, we've got these auto calling system and 70% of the causes handled by humans sitting in.

As you may recall or you may be aware of it since you're also deeply into it with Amazon Go stores, right? Like it was all supposed to be vision programming and all that stuff and everyone was like, how is Amazon figuring it out? You can just walk in, grab your stuff and walk out. You didn't have to bring a credit card and all that stuff. Like, you know, just grab whatever, put it back and all that stuff and it'll accurately calculate in your Amazon account gets charged. And it turned out it was just a bunch of Indians in centers watching the videos and then putting stuff in your cards, right? And two weeks ago there's another company, there's another AI tech company. When it comes to like AI chatting that came about, it just turned out that it weren't AI agents, it was big centers in India. Humans just talking, right? And these companies raised like tens if not hundreds of millions of dollars on that pre tax. It's just that in reality with FOMO and all this stuff, the due diligence is weak, right? Like I've raised money, right? 80% of the money raised is just based on. I like this guy, right? Saying all the right side. I like this deck then and fomo, right? Like, oh man, I've got my partner over here or I've got this company that just was able to get in early on this like you know, $12 million Series A around. On this particular company, we're kind of the ones that are missing out. So we've got to like deploy funds and we've got an AI fund. So we got to find AI companies, right? And somebody comes out as, as I can play that game, I'm telling you, I can absolutely tomorrow go and raise 6 million bucks, right?

So there's, there's definitely certain level of hype in there.

So is it. I, so I've always got a sense that there's some hype around this. To the extent to which it's hype and versus reality, it's, it's been difficult to sort of decipher that. Is it safe to say though that if I, let's say I get two or three AI, AI quote unquote, AI based applications in my business, maybe it's around submitting EOBs, maybe it's around reading radiographs, maybe, you know, it's, there's the verification. We're going to talk about verific here pretty soon. Even if some of these aren't there yet, so to speak, with full sort of automation and real AI neural networking going on to give these accurate customized responses to the prompts, whatever those prompts are. Do you think it's still a good idea though to be adopting some of these even if they're not there yet? It may not be saving you time and it's actually costing you some money, but as it continues to develop and it's feeding your data into it, are you going to be better off in a year or two than those who wait until it is more market ready?

No, I think, I think it's time, right? Like it's time because you also want to be ahead of the curve a little bit, right? It's like that you, you know, it's, it's, you know, you gotta be where the puck is going type of thing, right? So it's time, right? Like I think, you know, somebody said, was it time, you know, a year ago, probably not six months ago, probably not three months ago, maybe today? Yeah, it's time, right? Like it's time that you start adopting this because clearly that's the future and the tech moves faster and there's a exponential curve to it, right? Like just in humanity, right? Like 200 years ago of the first 10,000 years of humans weren't like very, you know what I'm saying? All of a sudden the curve is sharper and sharper and it'll continue to be, right? Especially with genuine AI that can do so many different things. Now I can tell you that my wife uses Chat GPT like it's, it's like her best friend, right? Every single thing is just a verbal conversation with Chat GPT. I think ChatGPT is a therapist. Chat GPT is her information, you know, and so, and I, I get. When it comes to Chat GPT and its language skills, I am beyond belief impressed. It's just, you know, I'm Indian, so there's a, you know, we speak a language called Hindi. And I'm able to have conversations with this in the way that it speaks. It's so fluent and it's so precise and it's mind boggling to me, right? Like, you know, and we say that, hey, write up a poem for this thing and it can spit out a poem, all that stuff. And the creativity displays. It's true. It's real, right? Legal. Many applications. I can tell you we have a lawyer. I, my, my. I do not utilize my lawyer as much. Nearly as much in fact. You know, we've got all these documents. If I have a question, I just feed it to the thing and I'll say, hey, tell me about this thing. And it gives me a very quick answer that would have cost me 1200 bucks because the lawyer would spend four hours, you know, reorienting and all that stuff, right? So undoubtedly great applications. I think it's probably one of those like pick and choose and know your level of tolerance, right? There are offices that tell me that they've used AI receptionist very successfully in their practice, right? And I have called the practice to test it out and it does not meet my threshold. Right? Like there's a very popular person who said, well, you know, if you ever want to see how the, this AI thing works in my office, just call this number, right? And a few days ago I did and I was underwhelmed by the experience and I was like, yeah, it doesn't work for my practice. Right. Like so. But it's one of those things that should get there sooner and the more you adopt it and the more you put into it, the better results you get out of it.

Like I said, if someone said what is the right time to start implementing, adopting this stuff, I would say it's right now with certain cautiousness and in, you know, one of those trust but verify types of, you know, mentality.

How much do you think the, the new technology of AI that comes out will be these from scratch new AI softwares and how much of it are going to be legacy software companies that layer in AI features? I mean look at practice management softwares, of course, Eaglesoft, Dentrix, Curve, Open Dental, all of these. How much do you think they're going to layer in their own AI? So you don't necessarily have to plug in outside AI?

Yeah, I don't. You know, again, like it's very context driven, right. If somebody says to me today that we're building a purely AI based practice management system, right. My first question would be why? Because the, I don't know, the cost benefit analysis on that is, is there, right. Like there are certain things that you're just better off writing the code for, right? Like now the, the area where this probably good application for AI legitimately is to just have AI write that code for you, right? Like there are so many AI coding applications out there which, where, you know, you can just have AI write the code that you want, right? But it's just legacy code is just written by a, you know, like a Gemini or something like that, right. Like, and so, and I think I'd see more and more of that, right? Like testing and everything, right? There's AI agents that are coming and testing code and all this stuff. We use that, right?

But when it comes to like purely AI based, like using one of the like, you know, chat, GPT type, plugin operations to handle the workflow for say like patient engagement and recalls and reminders, it's all, etc, yeah, I would probably just say there's no need for it, you know, I think you can write it, the code, especially if you write an AI thing to write the code faster and better probably.

So is it safe to say that you don't foresee a time or a. Yeah, a time or a context or a need where a dentist would custom build their own agents? Because I mean Chat GPT allows you to create your own AI bots, your own AI agents. You can, you can feed it data, you can try to train it and then you can plug it into some process that you have in your business. And I'm looking at my own business which is a little bit of an anomaly because I'm a CPA financial planning investment firm all in one. Personal financial planning, business financial planning. And then of course we do accounting, tax and payroll and it's many softwares. There's a lot of CPA firms out there, there's a lot of wealth management firms out there. There's not one that does all of that for specifically Dennis. So I have a very unique business model in that sense. And so I could see they're not being off the shelf AI specifically for a lot of my processes. There will be for the accounting and there will be for investments but a lot of mine know and so I, I'm creating a, a tech committee here at my company and I have a consultant who will come in and we're just going to map out our processes in our tech stack and say where are the bottlenecks in the flow here and where can we create more capacity for advisors to have more facetime with their clients to really address the human side of money? But in a dental office, there are definitely unique dental offices with maybe their own brand and style. But for the most part a dental office is, they're doing sort of the consistent same things across offices. So is it safe to say that dentists will probably never need to create their own custom bots and, and they'll always just use off the shelf products?

I would think so. You know, I think, let me say this but I think by the time that they do build these bots, right. There are companies working on the problems. Right. Like the, the and the, the company will probably come with a more mature solution than the dentist is using themselves and that might be the right one, might even be more cost effective or maintenance free, et cetera. Right. Like AI Notes is another one, another application that's, that's popular that I'm seeing a lot of companies being used. I have a friend that's actually working on an AI based AI Notes company, right. Where they just put microphones in every single operatory. Right. And it's just creating your notes automatically and putting that in 100%, a very, very good use of AI.

I feel like that's a really low hanging fruit. I mean I had two financial planning meetings today and I haven't adopted that Yet. And so guess who's typing out my notes for 30 minutes after the meeting? I am. And then I take my notes and I put in the action items in a separate software which emails the clients their homework assignments and my accounting and tax and payroll teams their homework assignments. And. But this is a very low hanging fruit because it is extremely good at taking these notes, cataloging them in a certain way and then creating the action items. And I did ask a few of my guests if they're using AI notes, if there's AI notes out there. I think I asked Overjet if they have AI notes and they, they don't have AI notes yet. So I don't know if there's any off the shelf AI notes. I suspect if you put the microphone and you use Firefly or there's plenty of these AI note takers, if it could do it. But it's not going to auto transfer the information to the right places in the practice management software. That's why I'm like, well, can Eaglesoft and Dentrix build in their own AI note taker in the clinical notes? I would think so, yeah.

I think that's what they'll do. I think either that or third party companies that will build it and then an Eagle Soft and Dendrix will just buy them out. Right. Like that's kind of the normal progression. Right. Some of the smaller companies with two founders are just so much more nimble than, you know, Dendrix. Allocating money, getting blah, blah, blah. Right. Takes a long time. And so, yeah, there's some things like you said are low hanging fruits, right? I know it's probably low hanging fruit. There's a lot of stuff, right. Like I can tell you that I have used ChatGPT for X, Y and Z. We have an orthodontist and you know, we didn't really have a very, very good Invisalign process in a manual. Right. And my, my brain was like, wait, we're not special here, right? Like we're talking about Invisalign, we're talking about a dental office. Nothing unique about it, Right? Right. And I just went in ChatGPT and I just said, this is the orthodontist name, this is our coordinator. This is what they do, this is how they do it. Make me a manual, right. And it just spit one out right? Now you just tweak it a little bit, right. And ask it to give it changes, etc. And it did a really, really good job of building something out. So I think there's definitely Many, many areas there. But I also think that, you know, the, the. I am a bigger believer now in like, just focusing so deeply on your core competency. Right. And letting the other experts. Right. In their core competency handle. And it costs money. But I still believe that instead of trying to save money and trying to do something yourself, you know, if you're just laser focused, zoned on just like increasing your dental collections and all that stuff, right? Like, and that's just a better use of your time and your energy, et cetera. Unless you just are passionate about it, right? Like, there's some people that are like, whoa, wait, that's my work. This is my hobby. And I want to, you know, and I want to do some things over here. But, you know, I would, I would think that, you know, around the country there's plenty and plenty of companies that we haven't even seen pop up yet, right? We've seen a lot. But, you know, for every company that pops up, we know there's 10 companies that are still working in the background on coming up with their first product or the minimum viable product product or the beta and all that stuff, right? And then it's just typically a tsunami. And we're starting to see that tsunami happen. Right. Like, I was at the DEO conference last weekend. It was just a remarkable number of vendors and remarkable number of vendors that have AI, Right. Like, so, yeah.

Well, let's jump into to your, your product, Verific, and why don't you give us a little showcase of what Verific does. To what extent is it introducing AI and how's it helping dentists?

Yeah. So insurance certification is the one missing piece, right? Like there are 20 companies in the space in what's called a patient engagement platform. Right? Like the full life cycle, whether it's like marketing, online scheduling, you know, new patient forms, you know, medical updates, et cetera, treatment presentations, reminders, appointment confirmations, you know, post treatment documentation and all that stuff, right? Like, all of that stuff is done. There's one missing piece to it, which is insurance verifications. Right? And literally you've got 20 companies. They all started from different places. Some companies started from being analytics and then they added scheduling and they added. Some companies started with treatment presentation and they added X, Y and Z. Nobody's been able to crack this problem because of the complexity of this problem. So we just took a different approach and we just said, we're not going to solve this problem 60, 70, 80% of the way. We're just going to solve it 97% of the way, not always in 97% of the way, because it's not really 8020 rule, it's really 8017 rule. And the last 33% of any system never gets done and the world just spins with things broken here and there. Right. And so we will just zone in. And so that's what our goal is. And we're the company that's further ahead with everyone. Right. We have the maximum amount of data there is. So what we did is we built a small software called VeriSmart. We started as a human company, right, which is centers out in India and Philippines and people will just make phone calls on your behalf to get this data. But then we just put up built a platform where a dental office can go ahead and customize the breakdown form. Right. This is all in context of an insurance verification. You customize it right there and then when we do an instant verification, it goes to the multiple sources, like electronically pulling the data, the portals or what we have built, which is the magic of Verismart, is a very big crowdsourced master database. Every single time we make a phone call, it goes into a master database we used across all our clients. That's why the more clients we get, the better data it becomes for everyone. And so every new client that comes on our platform is seeing far more greater value than some of the early adopters did when the data was sparse. Right. And so, and then what we did where we utilize AI quite a bit is the transformation to our standardized format. Right. In order for it to make sense to a particular dental office, it needs to read exactly every single time in the way they customize. Right. This also helps us to be able to integrate it back into a practice management software. So the data goes into it. So right now we're integrated with Open Dental, Dentrix and Eaglesoft. Right. And then more to follow. Right. We've got a couple of cloud platforms that we're starting to begin the integrations with. But the whole thing is that the very smart software goes to all these data sources, then utilizes the AI transformation to go ahead and transforms it into a standardized data, and then we can tell it what is missing from what you were looking for. And then you have the ability to either outsource it, which is to have one of our human beings call to get the missing data. Right. Or you don't really need all the data for everything, so you can just take import the data as is because that's more than enough for that particular patient, right?

Yep. So you used the word crowdsourcing, when I think of crowdsourcing, I think of companies like Upwork or fiverr or 99designs, where you have people really all over the world who offer a service and you have a website which they all come and sort of advertise themselves onto. And if I need a somebody to prepare a PowerPoint or if I need somebody to create a graphic, I can go to one of these places and I can put a bid out there and I'll get 10 or 15 different service providers want to provide a very inexpensive and what I found, very good quality on this. So I think a lot of our listeners will have heard of the term crowdsourcing. Now I've, now I haven't heard it yet in the term of AI, but it makes sense. What you're saying is you're crowdsourcing all of your users data. Now you're not, I'm sure there's privacy around the data and all of that, but you're bringing it all into one sort of place, one, one brain, one AI model. And then therefore every, every new dentist that plugs into that brain is going to have a better, accurate, more accurate success with your software. So, and so you're crowdsourcing all of that information. Instead of one dental office, it's 400 dental offices or a thousand dental offices. And then therefore the AI gets better and better and better because of the amount of data. And this is like, you know, Waymo is trying to constantly collect data and data and data so that these cars have enough data that they can spot any small little thing so that they, you know, turn left or they, they stop or whatever that is, that, that, that whole concept there. So my, my question is this then on this subject of, of crowdsourcing data in the world of software development. For one of my software is called Practice Orbit. It's a marketplace platform for buying and selling of dental practices. And in these marketplace platforms like Amazon or Airbnb or Uber, OpenTable, et cetera, the ones who get there first, who have enough people on both sides of the transaction make that software or that place, the dominant place that you're going to go because it has all the inventory. If I'm going to go and rent a room, I'm only going to go to Airbnb because it has all the inventory. If I'm going to list my room for hosting, I'm going to go on Airbnb because it has the other side of the inventory, people who are going to want it, it owns that space and that's called the network effect. You get enough people on both sides that it just sort of blows up on its own. YouTube blew up at one point, it was almost dying. And then it hit the network effect and it just blew up. Is there going to be a similar concept here with AI companies that the first ones end and the ones who build the biggest, largest models because of the amount of data they collect are going to be the ones who dominate because their product will simply and always be better than newer entrants that don't have the data.

Yeah. So in my space, the that's 100% true, right? Like, we needed to be first, we needed to get the network effect, right? That's the word that I use as well, that, you know, you hit some critical mass. And, you know, I fully expect that the trajectory of our company is going to be like this. And then all of a sudden, right? Like, and people in my company know that, you know, there's going to be a time very, very soon. Right. And we control it a little bit. Right. Like there's tweaks that we have to do. Because you also don't want to go too early because we deal with data and we don't want it to be bad data before it gets too popular. But I 100% foresee that my company is going to be, as soon as we dial it up a little bit more, it's going to go like this and we have an internal project where we need to know how much can we handle. Right. Like, my one concern is that there's going to be a time that we are going to start growing too fast and things are going to fall infrastructure wise. Right. So we need to be ready to be able to like, scale from, you know, thousand clients to 5,000 clients. Right. In a matter of a few months. Right. Like, can we do that?

So in my space, that's true. But there's one caveat that I always will add as well. It's not enough to be first. It's also very important be very, very good at it or the best at it. Because there are, as you know, with Facebook, it completely used the network effect. But there were other companies. There were MySpace and Friendster and High5 and Orkut and all these companies that were just building the same kind of stuff. And Facebook was just best at it as well. But early enough, right? So I think it's one of those things, you don't have to be first, you have to be early. Right. You have to have the best platform. And then of course, all the other things around it. And I think it's definitely, most industries are going to be like hero or zero, where you have like a, you know, it's never winner takes all. It's winner takes most of it. Right. And you know, we definitely keep going more and more towards winner takes most of it because of the way the world is structured. Right. Like big gets bigger and bigger. Right. Because it just now has more and more advantages. Right. Like that's why, you know, it's, it's sometimes takes way longer for a one billion dollar company to become a two billion dollar company than it is for a one trillion dollar company to become a two trillion dollar company. Right.

Like, so, yeah, it, it, it reminds me of, you know, that the phrase that data is the new oil. I feel like it's all the more applicable here. In AI, the data is like oil. The companies that have the data are going to have smarter AI models and the smarter AI models are simply going to do the job better than other companies. And so the, the book that I was trying to think of, I think it's called Blitzscaling by Reid hoffman, founder of LinkedIn and you know, pretty well known name there in Silicon Valley. And in Blitzscaling he talks about being the first one to market, but I think that he does talk about quality in the book too. The ones who do it well are the ones that launch early, but they're also extremely intentional about getting feedback, user feedback, and they're iterating it very quickly to get better and better and better. And because they're one of the earliest ones in the marketplace, they end up sort of dominating that, that space. It'll be curious to see five years from now. It's almost like when dot coms came out. Like everybody had a.com, but there were a few winners, obviously, Amazon being one of them who really came out of that. It'll be interesting to see if that plays itself out in AI that there's going to be so many entrants and there's so much PE money and venture capital money trying to find a home in AI right now that money's just like, like you said earlier, if you throw a pitch deck together using AI in five minutes and you put AI all over those, the slide deck, you're probably going to be able to get 5 million on something. But it's the ones who actually build a substantive valuable product, other ones who will win in this, in this space.

There's a time when the rubber hits the road, right? And then that's where all the management and the leadership and all that stuff counts, right? In, in. You know, when I started my other company, PPO Profits, we were definitely a new entrant in a space that had many, many much larger companies, right. And I used to just have this like, expression that you don't have to be first, you have to be best. The best will become first in time. Right? Like, just do all the right things, right. You know, it helps to be first. Right? No doubt. But I'm, I'm, you know, I will say this, like when I applied some of the lessons to myself as well, right? Like, I am the founder and CEO of Horrific and we are what we're talking about. It is like, so applicable to us because for us, data is king. I am currently on a quest to get the maximum amount of data as quickly as possible. Right? That's, that's our game plan. Like, let's just get the data first, okay? We'll grow later, right? And there'll be a time for the growth because data is the new world. Once we control the data, we've got so many different applications for that, right. So there's a little bit of me playing the long game, maybe to a little bit of the frustration of my investors, right? Like literally where we've got investors as well, and they see some of the other companies in this space come out there and make a bigger spot, Flash and all that stuff. And we're still kind of playing the long game a little bit, right? Because I'm very, very focused on we've got to have collect the data, make sure it's accurate or correct the data as well, and have the largest amount of the most accurate data and we will win.

This really, it really is similar to Marketplace, digital marketplace platforms like the ones I mentioned, because so many of those spend years losing a lot of money to hit the network effect, which in the AI world, the equivalent is you have so much data now that you're, that you're going to, you're just automatically going to scale fast because you have the data. In Marketplace platforms, when you have buyers, enough buyers and sellers on both sides of the, of the platform, you're going to sort of own that space, but you lose a lot of money, spend a lot of money. I guess this is why there's so much money going into AI. So much investor equity money going into AI is because it takes a lot of money to get to that point where you have the data to truly be effective as an LLM model for that specific function applied Inside of a business. Let me end off on this final question. Another area in technology that we often have to decide as business owners is do I go with a suite of products that are all built from the same company that interact like, you know, like a care stack is coming out and it's got all these different things inside of it that historically were unbundled. So you got the suite approach and then you had the best, sometimes we call it the best in class unbundled approach. So for example, to give listeners a clear example, in my world, I used to use a software called Thomson Reuters and Thompson. Reuters had an accounting software, it had a tax software, and it had a payroll software and then it had a client portal. And because it was all designed and coded from the same company, everything integrated really smoothly. So when we're done with the accounting at the end of the month, we literally would click a button, it goes into their client portal, they get an email notifying them that the financial statements are ready, and now, and now the process is done. Well, I loved the integration of the suite model, the bundled model. The problem is, is it wasn't open to playing with outside software. So it didn't have an API. And so I was stuck using it and only it. And yet there were these amazing other softwares coming out that allow you to do analytics on the financial statement, for example, or do custom report building as another example. And so we ended up leaving Thomson Reuters about a year and a half ago and going to QuickBooks, going to Intuit. Intuit is the owner of QuickBooks, but we're still using the tax software from the old company. Now we use QuickBooks. We have a separate payroll company called Rippling that we use. And then we have a few other softwares that pull out of QuickBooks and we can use those softwares to create very customized reports. So now I'm in this best of class unbundled model. But it's, I love the specific ability to create a product that's very tailored. Now, however, it's created new problems in that getting data from one place like QuickBooks into the tax software. Now I have to download it and I have to upload it every time when we do a corporate tax return. So it creates some friction. Now in other areas there's this trade off there in AI, in your case, you're focusing really deep in this one area of insurance verification. I interviewed somebody else. I won't, I won't mention who they are, but they did some insurance verification. They did EOBs. I think they Had a little bit of stuff going on inside the clinical room, maybe some automated client communication and they're trying to build sort of AI all around this area. How important or valuable do you think it's going to be to focus in one specific function and just become so good at that one thing that no one will be able to compete as opposed to creating more of a bundled AI suite that integrates well, I genuinely.

Think you have to do both, believe it or not, I actually genuinely think you've got, you know, it's just that it doesn't, you don't take one thing all the way to 100% profession before you start something, right? Like right now we're starting claims, right? Like that's our next thing as well because we can't wait. You've got to almost have to like have some parallel things but you have to first learn to drive fast in a narrow lane, right? Like that's the first thing. And when you're like catching a lot of velocity, that's the time you say okay, we've got to start. Because you also understand a lot of principles, etc, right? Like you've got a team around it, you've got culture around it where you've got things to be really precise. Apple's a perfect example. It wants everything vertically integrated but everything it releases works really well, right? Like that's just the culture of the company, etc. So I, I do. You know, it's like, you know, if we build the best in class insurance verification product, right, there's still going to be reluctance from dental offices because in their mind I've got this software care stack as you use that example and I'm already paying a subscription and they've got some insurance verifications, right? And now it's just matter of like do I implement more? So one of the things that we have done is, you know, we, in our own platform, we also give others a third party API. So there's a cloud platform company that is now utilizing or consuming a third party API to build their own, right? And so that's the approach where okay, why don't we just become really, really good at this and let others use our data, right? And so this way we are just, you know, it's, they can have their own platform. The data for insurance verification is powered by verific, right? And so they kind of get to have the best of all worlds for the dental office. And then it's up to us to decide whether we also want to start building out more of the Modules, right. I do think that you've got to build every module to be really good, right? It doesn't have to be perfect, but really, really good, right? People have, you know, you can give them ordinary product for free and people don't like it, right? Like they'll get more frustrated. But you know, in reality, the world doesn't exactly run like you plan it, right? So typically what happens is you build one good product, right? A bigger company just kind of acquires you to integrate that into this thing. And, you know, so, so I can, you know, foresee where, you know, there are companies that say, hey, we love your data, right? But we don't want others to have this data and so why don't we become the big company, you know, because we are the only company that has access to your data. And those are the types of things happen. And you use the example of Intuit. It's perfect example, just buys companies left and right, right? Like.

Right.

Exactly. So, you know, who knows.

Where the industry goes in this case as well, right. Ideally you want a full lifecycle all built in, but there's many constraints, right? Time and money and all that stuff and focus. You just can't. If something takes 10 modules, right? You just can't snap your finger and build 10 modules. And typically what happens is one company starts with one module and starts adding another company, starts from a different direction, right? And they all kind of merge at one point in time, but each of them have their, you know, strengths and weaknesses. Which is also what the beauty of America is, right? Where that's why you've got, you know, 10 companies that can solve a problem at different price points and etc. Right? And that gives competition, keeps costs low and all that stuff, right?

Yeah, you know, I didn't think about, I didn't think about this, but how verific right now is the specific application is to do insurance fair verification, but you're building this LLM of data which then can be leveraged by others. And so it's possible that you convert from an application company into a straight data and LLM company, that others build their own layered applications on top of your LLM data. I feel like Elon Musk is essentially trying to do that. It's trying to create more of a, of a data company which then can be used by others. For example, last I heard, I have a, I have a Model 3 and you know, I, and I charge it. Of course, it's got a very specific charging handle and device on it as opposed to other companies. But I believe all of the companies out there are now getting behind his design. And so he's sort of opening it up to other companies to build these, their own quote unquote application or their cars around that specific thing. And it'll be interesting to see which companies ultimately emerge as the data companies in a way and open up through APIs for third party companies to build layers on top of that and which ones will focus on the application side of it.

Yeah. So Verific, I can definitely see us mostly being a data company, right. Because the thing that we give of value the most is data. And I'm a very, very big realist. Right. Like, I would love to say, yeah, we're going to have our own platform, which we do. Right. Like, but at the same time, you know, we also own a practice and there's this like death by subscription. Right? Like everything is 299amonth, but you add up 299amonth 10 times 3,000 bucks a month. Right. Like, and so I'm cognizant of that as well. And so, you know, for us, instead of us trying to compete with a bunch of other platforms, what we are really good at is the actual data surrounding insurance verifications. Right. And so I 100% foresee ourselves providing the data. That's why we built a third party API today. You know, a Dentrix or an eaglesoft can call us and say, hey, we'd like to get access to your data and we'll spin it up like this right now, do we just become the ubiquitous data provider for everyone? Or like I said, the other thing that I foresee is someone wanting to incorporate us into their life cycle. Because insurance verifications such an important part that people would switch practice management softwares just because of us. You know, there are literally clients that are like, if you are not integrated with this particular platform, we would not move to it.

Well, it's such a bottleneck. It's such a bottleneck in the process, right.

And it's an expensive problem, right. They have Suzy on the phone, cost 2,000 bucks a month getting all this data. Right. And it takes time and there's errors and all that stuff. And so there are, I'm telling you this, we have clients that are like, I want to go to cloud, but the only reason I'm not going there is because of UVoriginal. Right. And so, and that's why we're like, okay, we've got a third party API, the cloud provider that you want to go to have them connect to us, we'll share our data with them and we're right in the midst of that. So what we're talking about, you know, will the progression happen in the next month or two and three months, etc. Etc.

You know, which practice management softwares does Verific currently work with?

Open Dental, Dentrix and Eaglesoft. The big three.

Great. The big three. Got it. Well, Verific sounds like a. Definitely a best in class. This is you. You've really laser focused in this one area. It's a huge bottleneck. It sounds like an amazing technology. So encourage all my listeners to jump on. Is it verific.com? how can they get a hold of you or somebody at the company to, to explore a conversation?

It's Verific Biz B I Z and it's Verific spelled like the word terrific but with a V. Right.

And so two hours in there.

Yep, exactly. You know, I should ask, I should ask Chad gpd, how many hours are in the word verification? But you know, as soon as you go to the Verific Biz website, you'll see, you know, watch demo and just schedule a demo. Lean company, small company. I'll be the ugly mug that'll come over and show you the demo.

Okay, that sounds great. Well, I've got to say this has been a little bit more of a. On the technical side, but I, I think incredibly valuable though. And I, I'm really interested to see how the evolution of AI continues inside of dental offices. This was really insightful. And the reality, I think you're giving kind of a reality check on what is actually happening with AI versus what is the buzz or the hype around it. I can tell you're very grounded and that your software has done a lot of true deep research and effort to build out the data model that you have. So I suspect Verific is a very, very effective tool for dentists. Vivek, thank you for being on the show. Any final words you would like our listeners to hear about AI in dentistry as they consider adopting it?

No, I mean just enjoy the ride. Right? Like it's, it's, you know, it's the. I get a little philosophical sometimes, right. But you know, I'm loving, you know, it's like this like Chinese expression, it says, may you live in interesting times and we live in interesting times, you know, So I love the ride. I love the new stuff that's come over. Right. And, and I'm so glad I'M born right now and not in, you know, 1915 or 1850 and have to go to war, you know, like, so we do.

These are absolutely unique times, and AI is one where we are all trying to forecast what is this thing going to do in the world and in our personal lives and in our business lives. So thanks for adding another layer of understanding to this whole concept of AI and giving us an introduction to verific.

Thank you so much. Thanks for having me.