[Jennifer] It’s a pleasure to be here again, David.
[David] One of things that a lot of people get confused about is they hear AI and it it’s made to look so simple to use, and there’s a lot more that goes on behind the curtain. And I want to get into understanding that to better appreciate it when people start using a product like Angel AI, they have a whole lot deeper insights into what makes that Angel AI special, but also the complexities that go into it. Start us off. There’s an article that you were that Pavan sent us. I had a chance of reading it, but if you could start there and talking about that.
[Jennifer] Sure. This article is about what Andrew Ang, the founding leader of the Google Brain team, his assessment of what an AGI is, right? and what he defines AGI or during a test of that, he proposes.
[David] Let’s talk about AGI. Yeah, but that’s initial. So for those that are new, AGI stands for?
[Jennifer]: So basically it’s nothing but artificial general intelligence, right? what does that mean, right? What he defines as an artificial general intelligence and the test is when an AI must successfully perform useful economic work and match a skilled professional over a multi-day experience. Right? So that is what a true AGI means and if you run Angel AI through that test, she will pass with flying colors. Why? She’s already doing a complete loan origination process that normally requires a symphony of people like loan processors, CPAs, expert loan underwriters, and loan closers, right? and Angel AI is able to complete all this autonomously and by doing this, we have successfully passed Mr. Ang’s AGI definition. So that’s what I wanted to start off this conversation with. and and to also tell you a little bit of how we have been able to accomplish this.
[David] That’s where I wanna go, is how have you been able to accomplish this? Because I believe you’re the only one that I know of any technology as in the mortgage space that truly meets this test.
[Jennifer] Yes, and as people think, right, it is not something I can get done overnight. Not even in like a year or two, right? It’s gonna take decades to do this and tons and tons and tons of data right? Our engineering of angel ai didn’t start last year, didn’t start ten years ago. It actually started thirty plus years ago. Pavan himself actually laid the foundation of angel ai and he actually coded and architected the system himself 20 plus years ago and over time it was carefully articulated. It was carefully engineered. And one thing I always tell Pavan, even in our conversations, is like you’re the most patient person when it comes to engineering and technology. He will have this vision and right. And even sometimes I would go say, my God, I have to get this release. I don’t what’s I I’ll just keep I’ll be baffled by that, right? But you’ll say, It’s okay. This is engineering. It is gonna take time. And and that kind of patience, these are all qualities that you need to have but from an engineering perspective you need to have someone who understands real math and engineering. Artificial intelligence is not just LLM. You just give it it’s not prompt engineering. You could just give questions and it answers. Great. But that is not real artificial intelligence. Real artificial intelligence requires very fundamental mathematical knowledge, very fundamental engineering knowledge to be able to achieve this. And this is why our engineering team is very effective. And what makes our engineering team effective is its eclectic mix of skills, experiences, and ways of thinking. What I mean by that is I have engineers who are researchers, lead researchers in the industry and I have engineers who are ex-military engineers, which brings in a lot of discipline and a lot of talent in there. And I have top of the class graduates and I have MIT graduates, right? And it’s not just that makes the team successful. I also have game programmers. not I’m trying I’m not trying to trivialize game programmers, by the way. All I’m saying is we also have game programmers. You need all kinds of engineers to build something of this nature. Because number one, Angel AI has complex math and engineering skills behind her that helps her drive these decisions. And two, you can you can have a state of the art backend technology, but unless it’s very simple to use, no one is going to use it and for that, you need the best UI. The best UI is always created by game engineers in my mind because number one games don’t have a how-to that comes with it right even kids can play with games adults can play with games it’s the broad spectrum so you have to have a UI that anybody can use always uses that that his grandma I’m sorry his mother uses nothing except WhatsApp on her phone but she also uses Angel AI that’s how simple it is and we had a borrower who is a grandma, but 70 plus years old and she was so happy how she was like telling one of our loan officers, hey, I was able to use AI to get my loan closed. And this is a grandma who was like 70 plus years old. So engineering is not just about the architecture behind it, which is also required, but it’s also making it user friendly to a broad spectrum of users. And we always say, number one, it has to be kindergarten simple. At the same time, even the all spectrum of users should be able to adapt to it. So that is what it takes.
[David] It’s a combination of the simplicity by just being able to literally talk into the system and giving it the most simple of prompts and it just keeps working with you until you get there. That’s why someone who’s a grandma that has n probably no computer knowledge can use a tool like this. Talk about this, how this is revolutionizing companies. There’s guys that we both know and work with, guys like Troy Kennedy over at Loan Works. He’s one of the innovators that elected to build his whole new company around Angel ai. this is having such a transformative impact on his business and also the in turn the industry, more and more people are looking at this, but it takes a significant commitment for
[Jennifer] for example, you know, Pavan has been doing this since he was what fourteen years old. and myself, I’ve been with the firm for 27 years now. My director of engineering has been here for 10 plus years. So it takes a lot of time, it takes a lot of persevering knowledge and it has to be done very carefully and it’s a steady stream of improvements. You can’t you can’t make it happen overnight. You gotta do cell by cell by cell by cell by cell, right? That’s what Pavan talks about, cell engineering. And that is why he even named his company Cellegence that owns Angel AI, right? So it like millions of cells that have come together over 30 plus years that’s what makes Angel AI so powerful. And as you said, Troy, one common quality that Pavan and Troy have is patience. Because he started working with Angel AI a few years ago. And along the time, he’s seen how Angel AI has grown. And he’s also seen how things were built step by step by step, even for running his own company. I mean we do the payroll for him, meaning Angel AI does the payroll for him. Angel AI he has like complex commission structures that are set up. Angel AI does all that for him. So you don’t he doesn’t have any additional resources who are doing the commission structure that he has because he has multiple tiers based on hiring structures and various complex structures that he has. But all of that he worked with Angel AI to get it implemented. So now all he gets is a commission file in his inbox and voila, the data is right there in front of him. And he was able to work with Angel AI and build it over weeks, months.
[David] Well, it’s taking the commitment on all your parts. What’s really interesting is sometimes the most simplistic the simple systems, at least on the surf, they look so simple when you interact with them. I’m talking about how you interact with Angel AI. It seems so simple, but the complexity that goes into that, I’m interested in some of the those that are developing you the individuals that you’re using in the backrooms to develop this product. Give us more insights into it. I I’m really fascinated that you used gamers.
[Jennifer] Yeah, exactly, right? one interesting fact I would like to really tell here is one of two thousand applicants actually get hired in our filtering process. If two thousand applicants apply for this job, one person gets selected. And I personally interview them as well and I’m not looking for a person who has the domain knowledge or anything, right? All it takes is a lot of discipline, a lot of passion, and to be able to enjoy what you do. When you have all these things, then you can build what you want to build. Right. So
[David] How do you design you manage a lot of this most you’re primarily the one that’s making this all happen come together when you have so many various participants how do you bring them together on a common goal of what a common output goal
[Jennifer] See, once you have a vision set, right, and you find the passionate people to accomplish that vision, you will be very surprised. Like I’m gonna give you a real scenario. This morning we had a stand up scrum meeting for my engineering team. And one of our engineers, William, was showing one of the latest UI features that he worked on implementing and itself is so exciting because number one, he did this, he was able to accomplish what he wanted to accomplish. And in showing that, he had so much more pride in showing that to all of us and the rest of the team. And he not only just did what he had to do, he took a step back and he looked at what was already there. You know what? if I do XYZ, it’s gonna make it much more better than what it is right now. He actually improved what we already had and the end product is a brilliant UI. This actually happened this morning, literally three hours ago, in our Scrum call today. So this is exactly what it takes to build a successful product. It’s not something that happens like in a snap of a finger. And he’s been working on this for a couple of days now. People underestimate the value of UI. To have the right UI takes a lot of effort and energy. And and when I hire engineers, I always tell them there is no UI engineer, there’s no back-end engineer, there’s only one engineer. A good engineer will take care of everything end-to-end. And they will be able to build it so that it, as you said, a grandma can use it because that’s what they expect the product to be. They’re so proud of what you created, they’re so in love with what they’ve created that they can’t wait to showcase it to the whole wide world. It starts within the team, but then when that gets released in through QC process, they’re so proud of it. In fact, one of my other engineers, we invited him to one of our mindset summit events and one of our product launches and he came to me after that and told me, Jennifer, I know I created that credit boost service that we launched. I feel so proud to see the launch process of that particular service that I created. And to see someone using it, it only made me so much more happier. Right. That is the level of compassion and sorry, that’s the level of passion that that you have to have to create these products. you can’t have someone who’s just writing a code and getting it out of his desk. That’s not who we hire at all. We hire engineers who really care about what they did and who really take pride in doing what they have to do. Then managing the team is not difficult at all. I just let them go on their own and give me the best of what you have.
[David] You get it to get come together that’s the key I mean everyone going out and creating but they’re creating around a unified vision I get that but there’s still it’s it’s tying it and knitting it all together so that this is that simple how is that accomplished
[Jennifer]It’s I mean again, I might make it sound simple, but it’s not simple. Of course it is not simple. It takes a lot of work in the back end. and especially when it comes to angel AI, all you see is a chat window. But the amount of engineering that goes behind it in order to create it is a whole lot of work. I have had engineers who had a sleepless night. In fact, last night one of my engineers was like didn’t go to bed until this morning actually. I was even telling him, you know what? No, you don’t do that. but still, once you enjoy what you’re doing, time becomes irrelevant. Time goes to still, right? So you you just do what you have to do and if you’re enjoying it, then you will never look at the time. And and that’s what happens. And how they all come together. I don’t know how to answer that question. It just comes together because that’s how we’ve created it. That’s how we have set it up and that’s how they all work in tangent because everybody is so proud of what they create.
[David] Well, there is a sense of pride. It comes out when you and when I get to come attend the company parties like the Christmas party, things like that. People are genuinely honored to be a part of the Angel AI, the Celligence team. And while they they’re they don’t feel like cogs in the wheel. They seem like each one of them feels like they’re contri what is the key to developing that kind of culture inside the company, Jennifer.
[Jennifer] Two things. One is letting them take ownership of their own project and giving them the autonomy to make a decision on the project they’re working on so as I said earlier, it’s not like one person creates the front end, the other person creates the back end, and then someone else is testing it and someone else is releasing it. No. One person takes care of one simple thing. Like for example, one of the major features that we have is Credit Boost. If that project is assigned to an individual, he is first understanding what Credit Boost does. How is it even used in real life? Then he’s totally in line with the business use of it. Then he comes in, designs how he’s gonna do it. He architects a solution for that. He’s doing, or he or she is doing it themselves, right? And then Myself, my director of engineering, Pavan, we’re all here to offer any recommendations, suggestions. We meet with them to make sure that they’re not stuck and they keep going. That’s all we’re doing. Nothing is enforced upon anyone. They have a monopoly of their own thoughts. And once they do that, they automatically get that ownership. And now they’ve finished all this and they’ve created the product. They are actually responsible to coordinate with anybody that they need to coordinate to get that product live. And it doesn’t just stop there. They’re supposed to monitor the use of that product in live scenarios and make sure that that customer experience is stellar. And if it’s anything less than stellar, then they’re actually receiving that feedback directly from the customers and they’re actually improving upon it. So when you make someone do this end to end, they automatically are able to create wonders. That is something that we instill in everybody from day one. We don’t split a task saying, okay, you do one per piece of that puzzle. The person A does the one piece of it, person B does the next piece, person C, and then my director of engineering or someone goes and puts all these pieces together. No, it doesn’t work that way. That product is your product. You take full ownership of that. And it’s not easy to do that because every time my engineers come and tell me when I’m interviewing candidates, they will come and tell me, is this front-end related project or back-end related project? I said it’s a project, right? And another very common question they all come and ask me is like, what does 30-day success look like? What is 60-day success looks like? What is nine, six months, one year? No, you are given a product. That product defines your success. Right? It’s not if that product takes one year, I’m gonna wait one year to define your success. If that product takes one week, I will define your success in one week. So every product that you deliver defines your own success. So these are the kind of parameters that we instill in our engineers from day one, so that they understand that they are responsible for the project that they’re creating. And I know it’s not easy, it’s just a repetition. We just have to keep saying that again and again and again and again. And it’ll stick because they’re passionate engineer are not only just top of the class, they’re also very, very passionate about what they do.
[David] Right. And you’re also creating a sense of ownership, which is what gives them the pride that brings it all together. Jennifer, you’ve been in the mortgage business for 30, almost 40 years, if I recall correctly. I mean, you’ve been in for a very long time. Where do you see the industry going? I talked to Pavan about this, and I would like to get your perspective from an operations. You run operations at the company there for all these years. Where do you see the angel AI influence taking our industry? First of all, costs are gonna slash. You guys are the lowest cost originators.
[Jennifer] Costs are gonna slash, more than costs are going to clash, which slash, which is again benefiting the consumer. From a lender’s perspective, you need Angel AI for you to not have any buybacks. I I actually listened to one of your own podcasts that you did, how AI was actually has to be trained properly so that it doesn’t miss anything and which would otherwise result in buybacks and you know, you know how buybacks can affect a lender’s liquidity very much, right? So it is time that everybody understands that AI is not just a wrapper around an LLM, it is real training data. I still like the example of what your one of your hosts, I’m sorry, one of your attendees guest set which is basically I think this person was a minor and he was the buyer was supposed to receive some alimony or child support for a minor and unfortunately the AI didn’t take into consideration when I was listening to this that’s why they need angel AI because the advantage with Angel AI is remember she has been trained with 40 plus years of data that Sun West has originated loans with, right? So number one, I love that podcast. and number two, when I when I listen to that sentence of it, I was like, man, that person that lender needs Angel AI because Angel AI is very well trained on that scenario and many more scenarios. This is why it it’s just there’s no point in spending millions and millions and millions and billions in just building a wrapper or over an LLM because what they don’t have is that forty plus years of experience in the mortgage industry. So you need a good data set to be able to train your Angel AI or your AI. Your AI only is good as the amount of training that you do. It’s not the number of GPUs that makes it successful, it’s the amount of real-time data that makes it successful. And that’s why Angel AI is the only AI that can actually solve mortgage industry’s problems. Not only just that, I know in one of your previous podcasts, Pavan was talking about all the abuses that Angel AI was able to identify as well. She will never miss any of that. She will never miss a compliance issue. And I will be surprised why any lender would not want to use Angel AI or any broker would not want to use Angel AI because, as you said, not only is reducing the cost per loan, but is also creating compliant loans, and it’s also creating loans for borrowers who really have the ability to repay and not putting them in any kind of distress. So, I think that is what Angel AI is able to achieve and that is where the industry should be looking at AI as well.
[David] Any question if it works I encourage people to go to your angelai.com website you have tons of testimonials from realtors from builders from loan originators consumers that have used this product and it’s making such a difference in the lives of anybody that picks it up and does that. I want to get your thoughts on blockchain. Where are we and how this is fitting in? There you’re developing not only your nucleus but all these other things, the twin, the artificial twin that I mean we have the avatars that we’re talking about. Where and this creates up new opportunities and new pathways for which we in which we do the loans and how we communicate. And not that I get excited about one of the factors is we’re going to get a loaner engineers back to doing what loaner engineers are best at doing, and that’s answering the direct question of the consumer, but taking all the other busy processes out where there’s a lot of failure points. And it solves for those and gives them back the joy of doing the thing that’s getting more done and a lot more done.
[Jenifer] Absolutely, right? It’s one thing. I mean, one thing that everybody loves is spending time with family. And Angel AI definitely gives that time back to you. So that I mean, you we’ve seen originators being stressed and, you know, spending sleepless nights on the eve of a COE or something like that, right? But Angel AI takes away that stress and gives you your time back, which is the most precious thing that anybody can have, right? Number one. Number two, as you said, twin and blockchain technology. I mean, Angel AI is the only mortgage tech that has blockchain technology incorporated within her, right? And I know the future is going to be about blockchain and being able to have control over the data that you have the documents that you use, everything. You want to have a secure vault that can save all your information. And that information can be made available through generations in your mortgage doc, whether it be your mortgage documents, your will, your trust, whatever it is, right? And you will be able to put all this in one vault. And not only that, you talked about the twin. The twin, you are able to train a digital version of your own with personal information that only you have access to. All of your personal data will be very, very secure in your vault in the blockchain. I mean, what more secure ways you can have to save your own personal data? Right? So that’s why you need to be on Angel AI to enjoy the benefits of having your own time back, having a secure future and having a digital twin so that you could be in advance of everybody else around you. Right. So this is why we always say, Angel AI, I mean, of course, it has this warranted intelligence. You don’t have to worry about when she gives you an answer, we warrant that and all of that information. But having a stress-free time, being able to spend back with your family, who would not want that, right?
[David] Yeah, especially with the accuracy that’s there, which is going back to giving you the peace of mind that when it gives an answer, you can bank on it.
[Jennifer] Yep.
[David] It is so good to have you back on the podcast, Jennifer. I love having you here. I want to get into in the future some other aspects of what’s going on in the development of this. This is a complicated product topic, and there’d no one better to have you on to really just kind of explain some of the complexities that we have. We touched on it today. We’ll have you back again soon. Thank you so much for being here today.
[Jennifer] Thank you, David. It was a pleasure as always.
[David] Always a pleasure to have you on.
[Jennifer] Thank you.
