[Jennifer] David, it’s always a pleasure to be here. Thank you for having me again.
[David] It’s an honor. You guys have some exciting news. There’s been a new well if we have a press release we’re gonna talk about a little bit later, but there’s been a new patent granted to Celligence. I can’t wait to hear about it. Tell us about it.
[Jennifer] Sure. So the patent that we really we it’s a foundational patent for the transactional language model. we basically just got approved on that model. To speak a little bit about that model, we basically have an AI that gives a warranted intelligence. Okay, what is a warranted intelligence, right? And everybody’s wondering, okay, what is it? People call artificial intelligence and things like that. But what is what’s warranted intelligence? Let’s take mortgages for an example, right? Would you want to do a mortgage with someone or apply for a mortgage, someone who’s gonna come back and say, you know what, maybe I can give you maybe, for example, hypothetically $100,000. Maybe I can do this, maybe I can do that. and have a bunch of disclaimers saying that, hey, I’m gonna give you this. I’m gonna say that I’m gonna approve you for $100,000, but there’s a bunch of disclaimers. You know what? This is not really the final response, something might change. You know, that doesn’t give you the warm and fuzzy feeling that you need when you’re actually going and buying a home. When someone tells you I can give you $100,000, you really wanna be able to take that for granted and move on with the rest of your life, right?
[David] They were in confidence, yes, exactly. With incompetence.
[Jennifer] Yeah, exactly. So that’s what warranted intelligence is. When Angel AI tells you something that you are approved for so and so, you are approved. There is no surprises, there is no last-minute changes, there is nothing that’s gonna change that. And as you know,
[David] This was goes back to the one of the podcasts you and I did a number of months ago, which you talked about probabilistic versus deterministic. And I thought that really stuck with me because I thought all AI was deterministic. I thought I could depend on it. And then more we’re learning, not so much.
[Jennifer] If you see some of these very popular so-called AIs out there, you will see a message at the bottom, right? This is not like warranted response or this is not I use warranty because that’s the language that we use, but it’ll say don’t take this as an advice, don’t take this for granted. It has those disclaimers. This may not necessarily be true. All these come with those disclaimers saying, don’t trust this, don’t believe this, right? How do you expect me as a consumer to go and apply for a loan or use that AI for anything that I want to use it for in a real life in my day-to-day life? Right? So that this patent that we really want is actually warranting that no one can ever generate an AI that they can claim is warranted without violating this patent. That’s what
[David] That is huge. I mean, that is so significant with in the marketplace from your positioning standpoint, because you’re going to have the ability to give a warranted approval that no one else in the industry is going to be able to do. And there’s a lot of people talking about all the technology they’re bringing that. We’re l you know, AI, everything is the and this is gonna be a significant strategic advantage. how solid is the patent, something like this. I mean I know who your patent attorneys are and they are some of the best in the industry and certainly the best in the United States. So they’re one of the top top firms. So this isn’t a flimsy patent or poorly constructed. They they are very good at what they’re doing.
[Jennifer] Exactly. Nobody can claim that the idea is 100% accurate and warranted without violating this patent. And the beauty of this patent, David, it doesn’t just stick with mortgages, right? This is going to be applicable for any industry, right? In in in in fact, the paper that we the press release that we were talking about is about a paper that our lead data scientist actually wrote, which talks very clearly about how Angel AI is different in what she does, right? Can AI have an intuition? That’s what he’s going to be talking about. And he clearly takes talks about that in his paper: that in high-risk industries, standard AIs doesn’t know what it doesn’t know. Right? That is the biggest disadvantage. When are you not able to give a warranted response? When you don’t know what you don’t know, then You’re never going to be able to give a warranted response, right? And this is why the right model to use in high stake domains is the one that delivers the ultimate balance between what they’re able to predict. In addition to that, they have to have a quantified certainty about Right? This is what they have to have a quantified certainty, which means when Angel AI gives you a response. She’s able to tell with 100% certainty that this is the guideline, right? That is exactly what an AI should be able to tell you. It’s not just predict, right? You’re not just predicting it. You need to be able to say that with 100% certainty. And that’s what this patent lays the foundation for. It actually is the patent, right? And again, as I said, it is not just for the mortgage industry. This is needed for any industrial application. And you will see in that paper that it was actually applied in one of the most critical industries, which is medical and especially pediatric care. Right. And it was actually run in that environment in a really one of the top hospitals, pediatric hospitals in the country and they were able to see the results very clearly. So it’s not just a mortgage industry, it’s any industry that needs certainty. Your life, your financial future, your assets, your lifetime of savings, all these need certainty. And these need warranties and an intelligence that’s going to be able to give you that is Angel AI and the patent that we won is exactly that.
[David] Yeah, especially when it comes to a home loan or medical, that’s a critical decision to operate on how to operate, all of those things. You know, sticking in the area of mortgage, which is our area of expertise, of course, it gives a consumer assurance when they know this is a warranted decision, whereas they work with your competitors or anyone else using an AI similar type product, not that anything is really similar, but it in this sense that they think it’s similar. This is going to be a significant advantage in the marketplace. Think about a consumer who wants to go and they got it they get a pre-approval letter. They’re in their mind, that’s a done deal. This letter means I have this loan. And we both know these pre-approval letters have been flawed and for decades. And we’ve all used them, we all had them out there. But the reality is now we for the first time have a decision. Let’s go back and review probabilistic versus deterministic, because really what you’ve done, you’ve established a deterministic, you have on 100% determined that when someone received an approval through agile AI, it is determined to be 100% a decision. Yeah, go ahead.
[Jennifer] I just want to add one more thing to that, right? It is not just issuing the decision. As you know, both FDIC and CFPB has this law where when you’re going to use an AI to make a decision, you need to be able to show how the decision was made.
[David] That’s where I wanted to go. That’s where I go. Yes. So it’s when Pavan and I were Yeah. By the way, Pavan put Jennifer up to doing this. This isn’t her favorite thing. Although you do such a good job on the podcast, you get downloaded so often and and I really enjoy having you on. but he was talking about that very thing. I want to expand on that. when you just expand on what we were you were just talking about, because this is so significant.
[Jennifer] Sure. So let me, I mean, Pavan, I’m gonna steal an example that Pavan always uses, right? so you go to a 10-year-old kid and you see a vase was broken, and you ask the kid, Hey, how was the vase broken? The kid is going to tell you his version of how the vase was broken. It may be true, it may not be true, we don’t know, right? We’re gonna take maybe his word for granted, maybe you’ll find the facts are different. Everything anything could happen, right? So that’s what we’re gonna avoid, right? But if you had a surveillance camera in that area or something that’s going to record each and every movement that had happened before and after the base was broken, you will have you would have collected all the data that led up to the base being broken, how it was broken and what happened after it was broken, right?
[David] And I think when more and more from a come a regulatory standpoint, we need to be able to prove this, especially as there’s AI is going to be at second guest. And quite frankly, I think we’re in it for some really challenging times for those that are using probabilistic AI, rather than deterministic, and because they’re not going to be able to give an answer to as to how they determine how they came to the result that they did, that they could or could not do the loan. Well, and one of the ones I mean Pavon uses the good example of a veterans that have been taken advantage of in when you look at the financing. there’s gonna be a class action lawsuit coming out in the industry about this. I mean, there’s just been some abuses in this. But when a disc company can go in and give you a roadmap to exactly every calculation that went into that, that is from a secure from a compliance standpoint, that is the gold standard now.
[Jennifer] That is the gold standard. Angel AI is able to give you that surveillance of how the decisions were made. That way there is no guesswork. And guesswork is the last thing anybody wants in highly critical domains.
[David] How is that if someone wants to look at that? How is that the guesswork, getting the guesswork? When someone looks at that decision, I love the example of the vase that Pavine gave me earlier that you just used. it is so important. How did something happen, whether it was a positive or negative or challenge in any way? We are all gonna we all have to determine how we came up with the determination, whether we’re underwriting manually or we’re using AI. We have to say, well, I use this calculation, that calculation, I use this income, and it’s all in the file, supposedly, in the electronic file, but in the file. So explain how the how you are able to, or angel AI is able to then document exactly how that decision on a progressive step-by-step basis. That is the real significant part of this.
[Jennifer] Yes. Correct. So as we’ve talked about in the past, right, it’s the cells, right? There’s millions and millions of cells that are working together in order to analyze each and every piece of information. Right. When these cells are acting together, all this information is logged. And she’s able to make the final result based on millions of cells all and that is how we’re able to trace the decision that she’s made. I mean, I mean, I you you probably know this very well. Back in the day, I’ve when I started doing mortgages or started working with Pavan, we used to have an underwriter. She always had this yellow notebook with her because she wrote the entire thought process while underwriting the file, right? That thought process is nothing but the cells that are working in our brain. And Angel AI is the same way. She is actually built of millions and millions of cells that is actually going through the data, comparing them against the guideline, and documenting every step of the way. And it’s logged in every step of the way as to how she was able to come to that result. And the underwriters back in the day uploaded that document into the system. Angel AI is just logging that information into the system, and that is why everything she gives a decision on, or everything she says, everything that she has given a result on is documented in the system. And that’s why when a regulator comes in and says, How did you make the decision on this file? Voila, here is the decision-making process, right? So it’s as simple.
[David] Yeah, here’s the printout step by step. this is again something that is putting you on a more secure foundation if you’re a lender using Angel AI. It’s just constantly looking for ways to be secure. This is all covered in the press release that’s been it has it been released yet? Is the press release out? Or is it already out? And so we’ll put a link to that press release, listeners, in the show notes of this, and we’re gonna be having Professor Air Di is it i he’s a PhD, is he that? Yeah. PhD the doctor Aron Brahma
[Jennifer] Aaron, yeah, he’s a he’s a professor. Yeah, Doctor Aaron Brahma, yeah.
[David] Brahma. we’ll be coming on with Pavon and I and we’re gonna be talking to him about this. The fact that you’re able to do this, why has it taken so long, Jennifer, to get to the point this that that this hasn’t been done before? Why?
[Jennifer] It’s you you I mean it’s amount of data, right? It it you have to have covered all possible scenarios. There’s 40 years of data that Sunwest has in the mortgage and lending industry, right? And being able to cover all the scenarios and every single one of them. It’s funny you bring this up. I would really like to quote this. But 13, 14 years ago, we were still building the AI. one of our wholesale brokers was very, very upset and he calls me. This is ridiculous. It’s like, okay, what is happening? He said, I gave an LOE, the borrower gave a loan letter of explanation saying that the source of the funds that was deposited, the cash deposit, was because of a garage sale. Okay, the AI looked at that letter of explanation and said, Hey, I do see that there’s a garage sale. Okay, is this going to affect the value of the property because the garage has been sold? Right? So the AI at that time was not trained to understand that garage sale is not necessarily the sale of the garage, but it is an activity that happens.
[David] That’s a great example.
[Jennifer] Why does it take this long? This is why. You have to have covered all possible scenarios. It’s not something that you can do overnight. And if you’ve not been in the mortgage business, you wouldn’t have come across these lot of explanations. How many times someone is going to source funds from the garage sale, right? That’s that doesn’t happen quite often, as much as we all, so many of them do garage sales, but sourcing that as the funds to close doesn’t happen often, right? it’s a combination of the words, but at the same time, you have to look at it from a little differently than how a human reads it versus the AI reading it. This is what it takes, right? You have to go through all possible examples. I mean, that’s a great example I’d like to use always. And this isn’t happened last week or last month or last year. Happened twelve years ago. Right? And that’s how long we’ve been training the AI. And forty years of data and now here here we are and we’re able to give warranted intelligence. Yes.
[David] Here you are. Well, congratulations, Jennifer and team, for pulling this off. It is such a significant milestone, not only for Celligence, Sun West, and that, but also for the industry, for the for many ways, decisions that are to be made, because now they’re being made with confidence. This is so true in mortgage because you got to have certainty that when you got an approval, you have an approval where you can follow through, sell the home that you’re selling, or make the necessary critical steps, many of those steps are irreversible. So if you do not have an approval, it creates a lot of la lawsuits. This is gonna put you guys again at a significant advantage. I’m just so thrilled that you take some time again, Jennifer, to come on and talk about this. I can’t wait to get over the with Professor aron and Pavan talking about more, get diving more into this. But this is an important announcement, listeners. There is now. A true warranted decision is one, and the patents are written, so anyone else who’s making those claims are gonna have to deal with that patent. And you go, like, well, why would someone patent something like that? Well, then one of the things I love most about Pavan, his old adody, he’s making angel AI available to you, the lending community, for free. That’s the thing. Is it’s not like there’s a huge greedy play here. Excuse me, that’s what Pavan and company is doing is making this available for free. So I’m excited for our industry. I’m excited for what you guys have accomplished. It’s a major milestone. Jennifer, thank you so much for coming on and being here with me today. Yeah.
[Jennifer] Of course, of course. These are the kind of patents that makes Angel AI’s valuation to be hundred and twenty billion. And guess what? This hundred and twenty billion valuation came in when there was like seventy patents. Today we have almost hundred and thirty five plus patents and the patent that we just got approved wasn’t even part of that hundred and twenty billion dollar valuation. So, you know, bigger numbers coming. did I hear that number? Over two hundred and thirty patents in this area? A hundred and thirty. Wow. Yeah, I was gonna say that. But still that’s almost double what it was when you got the valuation of a hundred and twenty billion.
[Jennifer] Hundred and thirty plus patents, sorry. Yeah. Hundred and thirty plus and the valuation was at Right. And and yes, exactly. Right. So why is it free?
[David] Well, there’s so much more. Is if people want to learn how I can learn more about this, check out Mindsetsummit.com and get out to one of the Mindset Summit conferences. These are inspirational. There, there’s a lot of people there that can help walk you through it. A lot of the Angel AI team. Many of the conferences, Angel AI is I I Angel AI has a booth there. You can learn more there. Or better yet, listeners, you can just go to ask Angeldot AI and start using the system. You’re a little cautious. You want to learn more before you dive into that part of the pool? Then go to angelai.com and read the testimonials. Read everything about tons of information, press releases, patents, all now it’s up there. It’s a way to learn more. Folks, the future is here now. And now we have one deterministic, guaranteed, confirmed warranted solution. That is this is a this is a major breakthrough. Jennifer, thanks so much for being here. Appreciate you.
[Jennifer]1David, thank you. It’s always a pleasure and honor to be here.
[David] It’s a good congratulations to the whole team because I know it’s a whole team that’s making this all happen. Kudos to you all. Thank you.
[Jennifer] Thank you.
