Episode Transcript
[00:00:00] Speaker A: Sa.
Welcome to Power CEOs, the truth behind the business. I'm Jen Godet, your fearless host, entrepreneur, investor and business and AI consultant. Why are we here? Because iron sharpens iron. And when we bring industry leaders, disruptors and experts together to share what's working in business and even what's not, we all have an opportunity to learn and grow. As a result, our businesses grow. And the ripple effect impacts not only our teams, their families, but also our communities in our world.
You're in for a treat. We've got a really great session planned for you today and I'm just going to dive right in. Listen, a bad medical outcome does not automatically mean somebody was negligent. But when something has gone terribly wrong, families want answers.
Physicians may feel immediately under attack. Attorneys have to determine whether there's a viable case that exists. And experts can face thousands of pages of records before anyone gets close to the truth.
What happens when we apply AI to one of the most complicated, emotional and high stakes decisions imaginable? Well, today we're going to dive deep into that distinction and into the business being built around it. Joining me today are two people coming at this problem from very different directions. Rodney Payton is a surgeon and medical legal expert who has served as an expert witness since 1983. He has reviewed more than 5,000 potential medical negligence cases.
And Jesse Tokbrik is a technology and AI entrepreneur who has spent decades building digital businesses. Together, they founded Peyton AI, an AI powered platform designed to help professionals triage and analyze potential medical negligence cases. Rodney, Jesse, welcome to the show.
[00:02:15] Speaker B: Well, thank you very much, Jan. Pleased to be here.
[00:02:18] Speaker A: I'm really excited about this topic and I'm going to start with you, Rodney. Let's get the terminology right before we touch the technology.
Someone has surgery, there's a complication, maybe the patient is harmed. That alone doesn't necessarily mean negligence. For the business leaders and investors who are not lawyers or physicians watching the show, where is the line between a bad outcome, a recognized complication, a medical error and medical negligence?
[00:02:46] Speaker B: Yeah, that is the prime distinction. I mean, they're all about bad outcomes. And bad outcomes can happen for a multitude of reasons. It may be that the patient wants particularly, well, to begin with and they've come in with, say, cancer. They already have an immune problem or they may have been smoking all their lives and bad chests. So you can have bad outcomes just because if you like the material you're working on is not the best and it doesn't heal properly, that's a bad outcome. It can Be a tragedy, but that is nobody's fault. And so that doesn't lead to medical negligence.
Now, the next thing, an error, some form of error can happen at any time.
None of us are perfect. You're not perfect, and I'm not perfect. So the question the court asks is, well, what was the actual standard? What was expected of a doctor in that position at that time? And an error is judged against what is what they term reasonable. Now, it's very hard to define what a reasonable doctor and a reasonable patient is, but that's what it is. Perfection. It is not. And so we have to start from the point of view that nobody's perfect, we're not perfect, and errors do happen.
I can give you an example, for instance, recently of someone who was given a tablet of penicillin, and they were supposed to be allergic to penicillin. Well, they developed a rash and they weren't very happy about it, but hit cheap for the day, and then they were perfectly okay the next day. Now, that would be an error in judgment, an error, given it was below the standard and it shouldn't have happened.
But then negligence itself has to have a third thing, and that is it has to have something that happened. But for this.
So for getting a rash or something, I mean, a lawyer would look at that and say, this is just de minimis. I mean, yes, it happened, it shouldn't have happened, but it did happen.
But it hasn't had much of a consequence. So you had to have a duty of care. And the standard which is reasonable, it has to have caused something. And whatever it's caused has got to have a significant damage potion to it. So all of those things have to be there for there to be a case in medical negligence. And a lot of times a patient doesn't know what's happened to them.
They quite genuinely feel they've been harmed in some way. They don't like what's happened. And sometimes doctors aren't the best at giving them the answers they wanted.
So they go to a lawyer. Now, the problem with that is a lawyer is not a doctor. And a lawyer is then saying, well, I don't know, because there's a legal term I've got to follow.
So that is why the whole process becomes mixed up. The lawyer's looking at legal terms, duty, you know, breach, causation and damage.
The patient, first of all, just wants to know what happened to me. Can somebody please explain in words that I understand what happened to me?
And a number of times that that's it. That's what they're looking for is an explanation where they were hurt or a loved one was hurt. Then at other times they're looking for justice. They feel that, you know, they've been harmed, that the care was substandard in some way and they've suffered more than a little because of it. And so that's where the justice comes in. So medical negligence is more than just a bad outcome. It's more than just an error or an error of judgment. It is that there was actual harm and long standing harm caused by it.
[00:06:20] Speaker A: That's very complicated. And it sounds like you've got so many stakeholders looking at the exact same situation through all of these different lenses. You have the family who wants to know what happened, or the patient. You have the doctor at the hospital who's like, did I actually do something wrong? I think I operate under standard care. The attorney saying, can this actually be proven? And do we have a case? And the insurer is evaluating exposure. And here you are the expert who has to determine what the evidence actually says.
So let's talk a little bit about that process. You know, we talk a lot in business about efficiency, but here delay is not just an operational problem. It can create financial cost, emotional cost, reputational cost, years of uncertainty. So walk me through what an inefficient system can do to the patient or family, the physician, the law firm, the insurer, and where is the most expensive waste he hiding today in this very human process?
[00:07:14] Speaker B: Yeah, that's a lot of one go. But let's, let's just parse that a little bit. You're absolutely right. The first thing we've got to understand is that justice is a balance. And the reason it's a balance is there's not just one person here. We have the patient who believes genuinely that they've been harmed. You then have a physician who gets a letter that says, we believe you've harmed our client.
And that physician still has to work. That physician still has to go in every day. He still has to make more calls. She might still have to operate.
And yet there's a cloud hanging over.
And since it takes time, and that is one of the biggest problems here, it takes time. The patient wants to know what happened and I want to know as soon as possible.
But that's not what a lawyer can do.
You have to then get all the notes and records. You then have to have an expert to look at those and determine whether those pillars are there.
And that can take months. And sometimes getting those records, it can take up to nine months to a year, just to get a record and to get a clear handle on what's happened. In other words, finding the evidence to back the truth.
And that then is a waiting time. And for a patient, I mean, they might wait a couple of weeks before they phone in again, say, anything happening or can you tell me anything? And the longer that goes on, the patient's trust begins to waver and say, wait a minute, there are two professionals here. Are they ganging up on me?
Am I not being told something? And the truth is there's nothing to tell.
And then on the other side of that, you've got the doctor. And the longer this drags on, and it can lead to suicide. And we take obstetrics and gyne cases, Sometimes a lot of midwives are even peripherally involved, never work again because they're so upset and so annoyed. There's no doctor I know in this earth that would get up in the morning and say, I think I'll harm somebody.
And they're gutted when it happens, absolutely gutted. And yet it takes this time to find out what happened. That's the normal process.
That's what the problem has been. The time to look at the thing carefully, to look at the evidence and determine whatever the patient's complaining of. Is there hard evidence in the record to back this up, or is there no evidence in the record or even evidence against what the patient thinks happened to them? And that takes very careful dissection, and that can take weeks or months under the normal manual way of doing it.
[00:09:50] Speaker A: Well, Rodney, it sounds like a very manual process. It's a lot of decision making, a lot of judgment. And so we're going to dive into. On the next segment, you know what this might look like, because the question becomes, how do you take more than 40 years of expert judgment living inside one person's head and turn it into technology?
And more importantly, should we. After the break, we are going to dive into how Peyton AI was created, why Rodney and Jesse chose each other as partners, and why relational capital may be one of the most undervalued assets in entrepreneurship. Before we go, though, be sure to join us in the Power CEOs Facebook group or connect with me on LinkedIn and tell me what business, AI, investment, or leadership questions you want me to dig into on the show. The show isn't just about collecting interesting ideas. It's about finding the ones you can actually use. We are going to be right back after these important messages.
[00:11:01] Speaker B: Foreign.
[00:11:08] Speaker A: Welcome back to power CEOs. The truth behind business. Stay connected to this show and every NOW Media TV favorite live or on demand, anytime, anywhere. Download the free Now Media TV app on Roku or iOS and unlock non stop bilingual programming in English and Spanish on the move. Prefer pots? Me too. Catch our podcast version at www.nowmedia.tv. from business and news to lifestyle, culture and more, we are streaming around the clock. Ready when you are. But let's get back to the topic du jour. AI models are becoming cheaper, faster, more available, which means simply having AI has become a lot less interesting.
The most important question to ask now is what do you know that everyone else doesn't? Welcome back to the show. I'm talking with Peyton AI founders Rodney Peyton and Jesse Toprik. And we're going to get started with Rodney Peyton. AI wasn't born because somebody sat around and said we should build an AI company.
What were you seeing repeatedly in that medical negligence work that made you think there had to be a better way?
[00:12:14] Speaker B: Well, the first thing we mentioned in last segment was about the time there's then the effort that is required. You know, with staffing, they've got to be paired. There's billable hours and the cost rockets up. And if you suddenly, you know, go from, well, here's a claim in, I'll get an expert. An expert can cost anywhere between 3 and 10,000 just to do a case, to look at it. And the costs are rocketing in court business.
So we had to start thinking, was there a faster, better and more efficient way of doing this?
And from my point of view, it came that laparoscopic or keyhole surgery came in in the 90s.
I was asked to think of a way, how can you train a surgeon who is working day in and day out and cannot go back to being a junior? How do you train that surgeon in a new technique such as laparoscopic? You all know about gallbladder surgery things. So I had to sit and think this out from square one for the College of Surgeons in London. And what came to me is the expert has in their mind a method of doing something they have trained in it. It then becomes subliminal. The frontal lobe switches out. It's all automatic going on. And so what you had to do is bring it to the fore again.
So in order to do this, we find that we have to bring the knowledge that an expert has, get it explicit, then put it in and try it out against various cases. And as you rightly said, I've done over 5,000 of these, try it against those cases without my judgment on them, and then add the judgment and train the AI as to how to look for those four pillars.
So that was a very difficult thing because once it started to do, few things came up. You had to have what's called human in the loop to look at them again and find what were the complications.
So that process that we used to train surgeons was the process we could use to train AI. And I was so lucky. I think Lynn was. My wife was having a podcast called Success is Never Accidental. And she interviewed Jesse Toprak, who is an amazing person, California. And I was just so pleased to have his acquaintance and to be able to work with him. And that's where it clicked when our conversation started that maybe we could use that process on a framework to introduce judgment and reasoning. Not just reading the papers, but putting judgment and reasoning in. And that's where Jesse came in.
[00:14:44] Speaker A: So, Jesse, I'm going to. I'm going to go ahead and ask you, like you've built tech businesses before you've exited them. There are thousands of easier AI products you could have built than one that involves not only law, but also medicine, protected information, professional liability, and jurisdictional differences.
Why this one?
[00:15:06] Speaker C: Great question, Jen. I think my answer will be because the hard problems are where the value is easy. AI products get copied in a weekend. This one requires medical methodology, legal understanding, privacy, engineering to the max, and jurisdictional nuance that I think a wall that most people won't climb commercially. It was simple. Firms and insurance were paying enormous sums in expert time and internal hours, as Rodney indicated, to find out that most cases don't qualify. And in fact, from the numbers we looked at, single digits, if you. If a law Office gets about 100 potential plaintiffs every month, they might take on three out of that. So there's a lot of time wasted in trying to qualify and look at the cases that they should be spending their time and energy on.
And basically anyone who can answer, is this worth pursuing faster and more consistently has a product that people will pay for.
[00:16:08] Speaker A: Yeah, I couldn't agree with you more. So let me ask you, listen, I've spent a lot of time around partnerships and investments. Founders routinely underestimate how much risk sits in choosing the wrong partner. Great idea. Wrong partner, dead company. Nothing gets off the ground. So let's talk about the two of you. What did each of you bring that the other could not manufacture with vibe, coding, capital or credentials? And what did you need to know about each other in order to say yes to building this? And I'll start with You, Jesse?
[00:16:39] Speaker C: Well, I've been blessed to meet, have met Rodney.
I've always been fascinated with both medicine and law and obviously last decade plus AI and this brought best of everything that I want to in life in one big problem that needs to be solved. But beyond that, Rodney brought something that I could not build, buy, or prompt. Four years of judgment and the credibility that comes with it. I brought the ability to turn it into a product, a business model, and a company that can scale. But without Rodney's expertise and his lifetime of experience, there will be no chance I will be able to pull this off on my own.
[00:17:17] Speaker A: And Rodney, same question to you. Why Jesse?
[00:17:21] Speaker B: Because of Jesse's background. He was speaking with Lynn and I realized what this guy has done in his life. Anybody that knows anything in the States will know exactly who he is. And the stuff that he has done is mind boggling. And so when we started to talk, it wasn't that we said, oh, yes, let's get together. I mean, I flew out. I was in the States, I flew out to meet him. We had long discussions and we found that our value system was the same. And that's what's important about a partnership. This isn't just a let's do something on the fly and get rid of it. This is long term and it's long term between us. And we find our values aligned. What we wanted to do, we want to help people. We wanted to help the patient primarily, but also the doctor. And we both felt that way. And if we could do this better, faster, more effectively, as we said before, that would work. So it was our value systems. And it was so important to find a partner that has the same system in whom you could trust over a number of years, because that's what it takes to bring this to the fore. And Jesse hasn't disappointed. I tell you, he works through the night like nobody I know. When we have clients and they come on to him and say they've got a problem, Jesse's on it right away. And through the night, he works on some way of solving it. And I've watched that firsthand and it's amazing. I'm just, you know, as a partner, he is absolutely brilliant. He knows what he's doing, I know what. I'm new and the two of us together has formed a great partnership.
[00:18:51] Speaker A: You know, I've seen it time and time again. I facilitated a lot of these tech meets, the expertise, partnerships, and it always wins when the values align. So I just want to say it louder for those in the Back of the room, make sure that you align in values, not just in the complementary skill sets. Because when things get tough, and they do, you need to make sure that you've got each other's back and that you have those values to back up on. So now I'm going to talk a little bit about relational capital and how you're going to the market through beta testing, through, you know, market testing, et cetera. Listen, you can raise money, you can hire developers, you can buy soft software, but trust takes years. Which, Rodney, you brought with your 40 plus years of experience.
How much did your existing credibility and your relationships matter in, in getting Peyton AI off the ground?
[00:19:40] Speaker B: Yes, it's matter. It's mattered a lot. I mean, obviously in my jurisdiction and in fact in several of the other. We're in Australia, United States, Canada, Europe, basically the southern Ireland and the uk And I have worked at some part in each one of those jurisdictions and I know people within them.
I have done LinkedIn, LinkedIn Live for many years.
We have a podcast called beyond the Gavel. A lot of lawyers are on that, particularly personal injury and med and egg lawyers that would tune in to beyond the Gavel, where we listen. You know, I mean, I'm an adjunct professor in Edith Kahn University in Perth as far as law is concerned. So I have law and medicine and education degrees and you put those together, but nobody wants to listen to you all the time. You mean use this old fogey coming on again. They want to hear other people. So that allowed me to bring other people onto the podcast.
Those people have been the nidus of whom we've used because we trust each other. They've been on the podcast with me and that they have then listened to see what we have to do. And they were our prime targets to get this off the ground because that's what we had to do, is beta test it, get them to help us to grow and develop so that relationship, that relationship capital is so vital.
[00:21:02] Speaker A: Yeah, I couldn't agree with you anymore. And so now, briefly, before we go to break, every AI company in a high stakes industry needs to answer clearly, where does the machine stop and the human become accountable? We're talking about where is the human in the loop right now? Patent AI says its assessments are intended to support, not replace professional judgment. How did you consider this? And where does human review actually come into play inside your workflow? And I think that Jesse's a great person to go for this one.
[00:21:32] Speaker B: Sure.
[00:21:33] Speaker C: Well, the AI organizes evidence and produces a structure assessment. It produces a very Detailed timeline and we provide references to every little claim that's made in the report. We produce what's called a merit score with reasoning behind it. A merit score basically looks at the facts of the case and only the facts of the case and tries to give an indication to legal professionals about whether you should be spending your time on this case or not.
What AI does not do, at least in our case, it does not decide whether to take a case. It does not give a medical opinion. It doesn't give legal advice. Every assessment goes to a professional who makes the call. Now we escalate to a human expert whenever the evidence is thin, contradictory, or the case sits at a boundary. Methodology flex. So I know that boundary isn't a limitation. It is the product our customers are buying. A tool a professional can stand behind. And that only works if the professional stays accountable.
[00:22:38] Speaker A: Yeah, I mean, at this point in time, we can't exactly hold the AI accountable yet, especially not in a court of law. So folks, if you're considering building an AI product in your own industry, don't start with what can I do?
Like, that's way too general. Start with what expensive problem do we understand better than everyone else? And then ask what can be standardized? What requires judgment? What happens if we're wrong? And who remains accountable? And build that human in the loop. If you are an attorney, an insurer, a healthcare organization, or a risk professional and you want to understand more what Peyton AI is building, you can learn more and request a demonstration@peyton AI.com I strongly recommend you do that and or get in touch with Rodney or Jesse on LinkedIn.
Building the product is just one problem. Convincing someone to trust it when money, medical reputations and real human lives may sit behind that decision is a completely different situation. After the break, we're going to talk about what went wrong, what surprised them, and the parts of building a high stakes AI company that don't show up in the pitch deck after these important messages.
A beautiful AI demo can fool you. It can look brilliant for five minutes and still fail utterly. When real customers, bad data, regulation, security, edge cases, and so much more show up, the real product starts where the demo ends. Welcome Back to power CEOs the truth behind the business. I'm Jen Gode and we're going behind the scenes of building Peyton AI with Rodney Payton and Jesse Toprek. Jesse, you've built these digital companies before. Talk to me about what part of building Peyton AI was harder than you expected.
[00:24:47] Speaker C: Well, one might assume that the technology was the tough part. It was Actually the easy part and that's perhaps because of the learnings over the last couple of decades. The what I noticed once I started actually having demos with the law firms, a lot of these people are working at the law firms. There are all kinds of people, different demographics, younger ones, older ones, the younger ones that understand AI and already at the forefront we actually met a couple of law firms where they are developing their own custom LLMs in house. They're really into it and there are other ones that want to do have nothing to do with it.
So we had to and we're still doing this to this day. And I'm sure we're going to have to keep doing it, really determine our behavior and how we going to handle a sales pitch, a demo during the meeting based on the openness, so to speak of the particular client. But I can say that sales and behavior change were harder than the technology.
Getting a firm with 15 year intake process that hasn't changed to insert a new step took far longer than planned in some cases. And validation consumed more time than in the engineering sprint. And I'm learning that you cannot rush proven a system is right in a field where being wrong has consequences.
[00:26:03] Speaker A: Yeah, I agree with you. And it's really funny because we all hear the problem and when it comes time to actually implement something new, especially a new technology and healthcare and in legal and finance, those tools, three specifically it takes a lot longer to prove that hey, this is not going to cause more harm than good. Because the first thing that every practitioner thinks about is is this going to damage my license, my reputation, my business. So I can definitely see where the sales process and, and that proof of concept was, was a little bit of a hurdle. Listen, one of the things I see when companies start implementing AI is they want to automate a process that they haven't necessarily defined fully or properly. Maybe they've missed a step, maybe they've missed a software that they use or maybe they've missed something that's just automatic to them in their brain. And AI can't fix that. It can't already know that step in your decision making process.
So when you started extracting Rodney's methodology, how much process mapping had to happen before the technology could actually properly function and make the same draw the same conclusions in the screenings as Rodney would.
[00:27:16] Speaker C: We spent months mapping Rodney's process before we wrote any meaningful code.
What we found throughout this learning process, I would say AI applied to an undefined process just produces undefined answers faster.
So you really gotta nail down what you're trying to accomplish here. And what I will say is that it was probably one of the hardest things to do at the beginning, at least not. I think we kind of nailed the process in terms of communication with Rodney and I when we have to solve problems. But hardest part was explaining what he does automatically.
He just knows it. Right. So basically what we're trying to do is download Rodney's consciousness and his expertise and make it methodological and provable.
And sometimes I'll ask him, well, how do you know? And he will say, basically, you have to reconstruct 20 years of pattern recognition into these rules.
I think that that made me a better technology lead, understanding how he thinks. But you don't fully understand your own judgment until you have to teach it to something that can't. Not alone. And that's. That was a challenge for Rodney. But after about a year of trying to deal with cases that are especially edge, cases that we haven't even thought about and really nailing them, I think we came a long way. Now I feel quite confident with the accuracy and the process and the speed of the process that we have.
[00:28:44] Speaker B: Yeah, that's very important. If I come in there just because I was speaking to one of the Lordship Justices the last week or so, and what he said was they'd gone in the senior law lords and in the UK and they tried AI to see if it could replicate their judgment, and they just took AI and stuffed it in there and said, you know, it's not really very accurate and it doesn't do this. And I said, how much time did you spend actually thinking about how your Lordships come up with judge judgments? How do you work behind that?
Oh, no, we just tried it.
And that was the problem.
That the knowledge that I had to bring forward and go into what we call nodal points and think of how I did that and then put it into AI, then try it out. Whereas others and very senior law personnel just sort of stuck it in there to see what happens and say it doesn't work.
That was the difficulty. And Jesse, you know, was very tight on how we know and how we think about how we do that to make sure that AI is properly trained. It's like you wouldn't take a junior law clerk and give him the biggest case, but they were just busy firing this into sort of a general AI. And you can't do that for many reasons. One, it breaks the privacy rules.
And Jesse will tell you we had to build this from the ground up. But also, if you don't train it, it's not going to happen. And what you said there by human in the loop, that has to be there at all times.
What we do is screening, and we have a human in the loop when we need it. And I have to rethink that. So all of those processes had to come together and we see when it's not being done properly. And then the master of the roles here said 3 months ago actually, that probably if law firms do not use AI in the future, they may be found to have been negligent in their practice to have legal malpractice. So he wasn't doing that as a threat, but he recognized that it's coming and that we have to deal with it and we have to train it properly. And that's exactly what Jesse made me do and how we work together on.
[00:30:57] Speaker A: Yeah. I want to highlight a couple of things for everybody who's watching. Number one, general AI gives you general, unspecific, and often wrong answers. It will hallucinate or do all of those things. And if you're using general AI, you shouldn't. You should be de identifying everything because you don't know how that privacy or that security has been handled with your vendors. So it's very important that you understand that. But what I want to do now is get more specific, Rodney, because you talked about privacy and you know, compliance is a big deal, hipaa, legal issues, all of these things.
So let's talk about that more specifically, because a lot of AI conversations get fuzzy. They're like, yes, I'm compliant. Yeah, that's not good enough in medicine. It's not good enough in legal or financial. So you guys are handling sensitive medical and legal information.
You're analyzing causation, as you talked about in the first segment. You're analyzing standards of care. And you know that the wrong answer has consequences.
So how did you guys think about security, privacy, hallucination, risk validation, and human oversight as you built this?
[00:32:03] Speaker B: Well, we started with that. I mean, Jesse started with a clean sheet and said this has got to be on standard. What a lot of people try to do is get an AI, as you said, and then add all these on top of it. And that does not work.
Jesse set about at the very bottom working out all the security standards, all the various acts around the world about patient privacy. And maybe Jesse would be the best to speak about how he did this and how he has made absolutely watertight certain that patients records are safe and secure.
[00:32:39] Speaker C: Sure. So, Jan, we're dealing with some of the most Sensitive topics probably anybody can that will have to deal with in their lifetime. You know we're talking about medicine and, and law.
So when we started with this project we went with compliance first and the challenge was that we were going to be live in multiple jurisdictions around the globe. All 50 states in the US, Canada, Australia, EU and UK.
And you know, even within, within Australia there are multiple different jurisdictions. So we have to be completely jurisdiction compliant. So that required HIPAA, GDPR, ISO. Currently we're working on or getting our ISO and SoC2 certifications amongst many, many others. But I will say just to make a long topic that can be written a doctoral thesis about short, I'll give four principles. Who's anybody looking at the topic like this? We're looking at medical legal topic 1, data stays in a controlled environment.
So that is a must in terms of. These are requirements for a lot of these ISO related certifications anyway. Two, it's never used to train public models. So the way we have interactions with the any kind of, you know, LLMs, etc. When we need to, we have BaaS and we also have, this is the also requirement for all the jurisdictions we operate in. The private data, if it's a UK patient has to stay in the uk. So all the servers where the data is stored as well as where the processing of the case triage is done has to be in the UK and same with wherever we're at. So it requires a lot of technical setup. If you imagine we're all in these 50 states and all of these countries, we have to have servers dedicated in each of these restrictions which is a technical nightmare what we solved. And then three, every conclusion is tied back to the source record so a user can click through and verify.
There is never a situation where an opinion of a patient or a family or even the attorney leaked into the analysis. The system is completely trained to ignore any subjective opinion and completely and only base the analyses on objective medical documentation.
And last one is if the system cannot point to an evidence, it cannot make the claim. It's very important.
It's designed to say insufficient evidence rather than guess. And the output is structured so it's not a persuasive prose. Because fluent language is the most dangerous thing in AI. A confidence sentence is not the same as confident finding. It's kind of like people confusing confidence with competence with people.
[00:35:27] Speaker B: Right.
[00:35:28] Speaker C: And we built the product so the user sees the evidence, not the eloquence.
[00:35:33] Speaker A: Thank you for that. And so just just to kind of recap a little bit, David Data sovereignty is a reality every country is moving towards on premises.
They want that data sovereignty. They don't want it leaving their borders. So something to definitely consider no matter what you're building. But before you add another feature to an AI product can create your one page failure map, where can the system go wrong? How are we going to know what happens next?
Who reviews it? What data can never leave an improved environment?
What decision will I never fully automate? And if you can't answer those questions before you start, you're not ready for sure to scale your technology.
If conversations like this are useful to you, make sure you join Power CEOs and connect with me on LinkedIn. Send me your hard questions so that I can get them answered on the air. Coming up, we're going where founders and investors really need to go.
Why now? What happens to traditional consulting and expert witness business models? Where is that actual enterprise value in patent AI? And if large companies are racing into legal AI, what keeps a specialized company from getting flattened? Coming up after these messages.
Welcome Back to power CEOs. Stay connected to this show and every now media favorite live or on demand, anytime, anywhere. Download the free app on Roku or iOS and unlock nonstop bilingual programming in English and Spanish. Prefer pods? Me too. Catch the podcast version at www.nowmedia.tv. from business and news to lifestyle, culture, finance, law and more, NOW Media is streaming around the clock. Ready when you are, folks. The market does not need another company whose competitive and advantage is we use AI. Pretty soon that's going to be the equivalent to saying we use email.
We all do it.
The harder question for founders and investors is what do you own, what do you know?
Who trusts you and why will you still matter when the tech changes again six months from today?
Jesse, why is this the right moment to build this company? Why not three years ago and why not wait for another five years?
[00:38:12] Speaker C: Well, frankly, at a personal level, I happen to have met Rodney a couple years ago and that sort of prompted everything. It wasn't something in my mind, but sometimes opportunities come to you and you have to make the best of it.
But obviously it's not going to apply to everybody. But I would say in general, for anybody in, you know, trying to get in the AI business, trying to figure out why this would be a good time to solve a specific problem. I would say three things converged for us.
The models finally got good enough to read a 10,000 page record accurately.
That wasn't really true three years ago. You could kind of get to there with the ocr capabilities. But now we're getting records.
You know, I think Ronnie looked at like a 12,000 page record from analysis with like, you know, really hard is, you know, doctor's handwriting, you know, notes on the edges of a paper. And then we also now are able to process raw DICOM medical imagery. These are like X rays and CT scans. You can just drag and drop all of these files right into Peyton AI and have it processed. We can even process a video of a surgery and just basically analyze it if there was any kind of missteps perhaps, and, you know, malpractice during the surgery. And was it done up to standard? So all of these weren't really feasible. If they were, they were very expensive. So that's technology catching up with what we need to do, number one. Number two, the buyer stopped asking should we use AI and started asking which one?
And as Rodney mentioned, around the globe, even in the most traditional circles, AI is becoming more of a no, you kind of need to use it. It's not even an option now. You've got to pick which tool to use.
So in that case, going to number three. Now the economics of expertise are shifting.
Clients are no willing to pay. They're no longer willing to pay for hours. They want to pay for answers. It's got to be actionable.
If we wait three years and the incumbents have already chosen their tools, this was our window.
[00:40:20] Speaker A: It makes perfect sense. Timing is a, a little bit of luck and a lot of, of. Of appropriate market timing. So, Rodney, I'm going to ask you, you've built your expertise, expertise over decades. Historically, that expertise gets monetized one hour, one report, one engagement at a time.
Does AI commoditize experts, or can it make the best expertise dramatically more valuable? Because now the thinking can scale. Talk to me about what's happening in that consulting environment.
[00:40:50] Speaker B: Well, the first thing is, if you take large companies getting, you know, hundreds of cases, thousands of cases a year, they're getting snowed. And even though some of the very big ones may have a doctor on staff or a nurse on staff who tries to strain string these, try screaming twelve and a half thousand words, you know, pages that is so difficult for any human being to look at, and those records are getting longer and longer and longer.
So there comes a time when you have to cut out and you have to decide, well, which ones are worth going for and which ones should be cut out of for justice, that it's not right to get a client to keep going for that length of time before somebody can get round to giving them the answer they're looking for.
So the first thing that we do is the screening. And that's what it's all about. It's that fundamental due diligence. First look at a case to see whether it has those four pillars, you know, duty, breach, damage and consequence. That's what you've got to look for.
If we can get those out of the way, then we've more time to concentrate on those that should be concentrated on. And as Jesse said, those are in single digits in terms of percentages.
So, no, the expert's not cut out at all. What it is is taking the load off and allowing them to concentrate on those that have merit. It's taking the load of the lawyer because their firms are also getting snowed in more regulation all the time that they have to deal with. And it's taking that load off them as well by saying, these are the ones that's worth spending your time on.
So AI is not replacing anybody. AI is AI powered. If you like medical information, patent, AI stays in the middle. It doesn't. Most AI tries to please you. It stays in the absolute middle. That evidence is either there or it's not there. As Jesse said, for every word that RAI comes out with, you can click on it and it will tell you what page it's in.
So you have that security, a trust but verify, I think it was Reagan came out with that one. But you have that security of knowing.
All right, this is a case worth doing for the following reasons, or we are here is the reason to give to the client as to why we're not taking it on. And we can get rid of that in a matter of sometimes days at the most, as opposed to four, five, and six months.
That frees the lawyer up, it frees the expert up. And you know, it actually at the end of the day, frees the courts up. So it's not replacing anybody. It is giving them powered assistance to move forward.
[00:43:36] Speaker A: I really like that. And, you know, what ends up happening is the cases that need to move forward are moving forward much more rapidly. So you're getting closure more rapidly instead of waiting sometimes decades to get an answer from some of these situations. So let me stay on that note, for the founder watching, who has deep expertise in an area, maybe it's finance, health care, insurance or something else. And they're thinking, I know a problem AI can solve.
Rodney, what would you tell the domain expert? Like, what do you wish someone would have told you before you started this process?
[00:44:10] Speaker B: I kind of Got it myself about 30 odd years ago when I realized that I do things automatically. Experts do things. That's why they're experts. That's why you're slick at what you do. Should it be running a television broadcast to operating it is subliminal. You do it, you don't have to think about it anymore. You have to become. That's unconscious competence. You have to drop back one to being consciously competent. You have to go back and look at everything you do and be prepared to question, be questioned deeply on what you do and why you do it. And that for some people is very difficult. They can't move back.
And so for anyone who wants to do this, it is a matter of taking that expertise and knowing how to bring it to the front of your brain to work it out into logical steps and reasoned steps and then to be able to transmit that to somebody else or something else, which is the AI. Jesse did that to me. I knew it was coming. It took a lot. And we're still doing it. We always meet new things. We're not perfect either. We always meet new things and we head them straight on. Nothing sits in the world of the unspoken. If we see another thing that's coming up, a different look, then we sit down and we work together to solve it for our clients. And some of the biggest companies that we've worked with, they give us these things all the time and we work with them to help them get closure. But that's a matter of thinking and looking after your client.
[00:45:45] Speaker A: You know, I'm gonna, I'm gonna have to wrap this because we're getting towards the end and I feel like we could have another couple of segments, unfortunately.
But for anybody watching who's considering an AI company, leave this question on your whiteboard. If the technology becomes cheaper, faster and available to everyone, why does my company become more valuable? If you don't have an answer, that's the strategy work. Technology can get you into the race, but it doesn't automatically give you a reason to win. Rodney. Jesse, for the attorneys, the healthcare organizations, insurers, risk professionals or potential partners who want to learn more about Peyton AI, where should they go?
[00:46:22] Speaker B: Well, they can go on the site. Jesse, you take that because you set the site.
[00:46:26] Speaker C: No worries, just Peyton AI.com is a website. You can go there and you can request a demonstration there. And both of Us are on LinkedIn. You can DM us on LinkedIn. If you're an attorney, insurer, healthcare organization, and tired of spending expert time to find out a case doesn't qualify.
We'd like to show you a better way.
[00:46:46] Speaker A: Thank you so much to both of you for bringing the value today, folks. The bigger lesson here isn't whether AI can perform more work. We already know that it can. The leadership question at stake is what work should be accelerated, what judgment should remain with the human and be protected? What expertise can be turned into an asset without compromising the trust that made it valuable in the first place? Place for all the founders out there, that's the business model question you need to ask. For the executives, it's the operating question for you to ask. And for the investors out there, this is a valuation and a risk question. Make sure you are getting to the nitty gritty in these questions and answering them in a way that will set you up for success. These distinctions are going to matter a whole lot more as AI gets better and matures. I'm Jane Gode. This is power CEOs, the truth behind the business. Unfortunately, all good things come to an end including, including this show. But you have action to take. If you want to learn more about Peyton AI definitely go to go to that website or reach out to Rodney Payton or Jesse Toprek on LinkedIn. They are very responsive.
But think about what we talked about each segment. Pick one question and answer it today. Not tomorrow and not next week, but today so that you can move powerfully forward and start that forward momentum in your business.
We'll be back same time, same station next week. Until then, win today, win this week and I'll see you next time.
