[David] Allen Pollack, let’s get over to you and talk about technology. Everybody is talking about AI. We’re starting to… we’re, it’s just, growing at a crescendo. You’re one of the top AI consultants in the space. Is this leveling off at any point, or are we just gonna continue to see-
[Allen] No. No, it’s not. The and you know what’s funny is there is a huge influx of people trying to code their own systems. I talked about this, Yeah … two weeks prior in, in another podcast that we’ve had. Everybody is now building their own CRM and their own, loan processing system. Th- these are not gonna work. If you’re listening and you’re trying to build something the ability to move something into the mortgage industry and be compliant and have secure data and access to systems requires a significant amount of knowledge. And also, not significant, but it also an amount of investment and cost. You’ve gotta have your servers. You can’t just launch these things on your local MacBook Air or a computer in your basement because, you built it in t- 20 minutes. These systems don’t scale. Automatically writes database queries. So anyways, David, we can get into more details, right? I’m going down a rabbit hole here. But yes, it’s everywhere. There’s great uses for it, and there’s some things that are continuing to evolve and heat up. Something that… and I’ll go off on this quick little, this little description, David. But 20 years ago, when I started Nylex with a partner. Yeah. I was his part- gen- another gentleman’s- There you go partner in New York Loan Exchange. We were really the first pricing engine. We changed the game in how you saw pricing- … and eligibility. And we created millions and millions of loops that ran in seconds. In order to get the rates in our system, we created bots that went out. This is before you had RPA and bots now. But we found a technology provider. The internet was kinda new at that time, too. I don’t know how we found it, but we did. And we created screen scraping bots. I remember Bank of America… Not Bank of America, rather Countrywide shut us off, and we had 30, 40 customers that used Countrywide correspondent rates. And they said, “You’re automating access to our system.” Eventually, the customers complained and they had to turn it back on. But anyone can now build a bot, is my point. Anyone can use Claude, Google Chrome browser route, and it can log into websites and scrape data, click drop-down boxes, and fill out forms. Oh, sure. Yeah. Yep. Yep. The problem is it’s not secure. Your data’s sitting out on the cloud, and it’s being preserved and stored. That’s a good point. So do not implement this into your daily workflow. Mortgage companies, by the way, the folks, for the executives or senior level managers and folks that are listening to our podcast- AI policies and technology policies, the training, those videos that you send out to employees and they skim through it and they answer five questions at the end and they’ve passed and you never do it again. They’ve gotta be updated. You’ve gotta… People need to be trained. You have a huge risk if you don’t. So anyways I just mentioned that quickly to answer your question, David. AI is it’s a great thing, but we just gotta be really careful of it. But to your example I’m gonna get into a little bit of the tech story and we’re gonna have some fun stuff that’s actually gonna talk about AI. So let’s do it if you’re ready. Let’s do it. So here’s an AI blooper of the week. This is The Telegraph. They’re a, publication that everybody knows. They left a ChatGPT instruction in a published article. So the UK’s Daily Telegraph published a major article about Trump’s meeting with China’s president, and they accidentally left an AI editing prompt visible in the middle of the article live on their website. Oh. The paragraph read, “To further divide the piece and maintain the authoritative broadsheet pace, here are two additional subheads.” That’s what they wrote as part of their prompt, and the prompt goes on beyond that. It was quietly removed after the publication was once, once people noticed it, rather. And what’s interesting is, this is a 150-year-old newspaper, and they let their AI ghostwrite geopolitics and forgot to delete the instructions. And we wonder why borrowers don’t trust automated decisions yet. Yeah. So- Great point … aI’s making its way everywhere, David. You d- who knows what video or picture you’re looking at anymore these days. It’s slop, is what people in the music industry call it. AI slop. So I’m not suggesting everything’s slop. I love AI. I have an AI company myself, but again to answer your question, it is everywhere. And then check this out, David. The Pope wrote an article or he said something about AI. Pope Leo issued the first major theological document of his time so far. Yeah. All about AI. It’s a 42,300-word document. It’s called Magnificent Humanity, and it’s about preserving the human person in the age of AI. What I wanna know is the 42,000-word document he wrote, did he use AI or not? I guess not, but just had to throw that in there. Yeah. The Pope presented it personally at the Vatican, which is unusual, and among all the speakers in the room was the co-founder of Anthropic, one of the leading AI companies, if you didn’t already know that. Yeah. The Pope’s words were, “Artificial intelligence needs to be disarmed.” So I thought that was quite interesting. Interesting. Yeah. Let’s move on. By the way, how cool would that be to be at the Vatican and listen to the Pope talk about AI? Yeah. So let’s talk about some things in mortgage. TD Bank they say that they went from 15 hours to three minutes. They deployed their first agentic AI model, cutting mortgage pre-approvals from 15 hours down to three minutes. It generates an application summary memo for underwriters. It handles document gathering and data verification all on its own, and it is a step one of a full end-to-end transformation they’re talking about from document submission all the way through to funding. It’s been live since January. It’s not a pilot, by the way. It’s in production. So they’re saying 15 hours down to three minutes. Next news article, David, is LoanLogics, company that I used to be a co-founder of. Oh, yeah. Let’s talk about, let’s talk about LoanLogics. We hired you many times. You were a part of many of our fun, fun endeavors, so let’s let’s talk more there. So they just launched what they call the LoanBeam Broker Portal, giving independent brokers access to automated income calculation technology that was previously only available to large lenders. Brokers can now upload files, submit loan orders. They can do everything including tie their shoes. Brokers are about to get 20% off the market. Brokers are about 20% off the market. This is from LoanLogics, by the way. 20% off the market, but have historically been shut out of enterprise-grade income tools. The gap just closes is what they’re saying, so very cool from LoanLogics. Let’s talk about MISMO. MISMO has just released what they call FRAME. That’s F-R-A-M-E in all capitals. It’s the Framework for Responsible AI in Mortgage Ecosystems, a practical AI governance toolkit built for the mortgage companies of all sizes. The key message is that even if a lender doesn’t own the AI model, they still own the outcome. By the way, National Mortgage had mentioned that. And what MISMO’s doing covers vendor oversight, specifically what data they use, what borrower data is retained, and how model changes get communicated, and whether fairness in the testing exercises exist. So I thought that was pretty interesting. Hats off to MISMO. And Attom, David, that’s A-T-T-O-M, is a different story. Attom rebuilds its AVM from scratch with AI. So they just launched a completely rebuilt AI first automated valuation model. That’s what AVM stands for, folks, if you’re new to our industry. Using 30-plus years of historical transactional data specifically designed because comp-based AVMs break down in low transaction markets. So the result is a 2.9% median error rate with more than 80% of valuations landing within 10% of the actual sales price across the 98 million properties. I’ll repeat that ’cause that’s a big number. 2.9% median error rate with more than 80% of valuations landing within 10% of actual sales price across 98 million properties. I’m just gonna say, holy shit. So I’ll leave it with that. Yeah. That’s- That’s Attom, A-T-T-O-M. And this is just a quick little bonus article, David, that I’ll throw in there. First American adds AI to title. They’ve just added an AI assistant directly to their title platform. It’s the last part of their mortgage process. It’s an AI native experience layer. And what they’re saying is the full stack is finally getting covered from end to end is is a key point about what they’re doing. So if you’re interested or you’re a First American customer, you wanna check that out, but they’ve added AI to their title services. Interesting. Outside of that, David, yes, AI is everywhere and everyone is using it
[David] Yep. It’s really- Including the Pope Yeah. Well- Including the Pope. Yeah, I’m sure he’s used it. You can’t say he hasn’t. Bill. Thus, he’s got that… Bill’s, if you guys can see the camera, guys, Bill’s got that shy sher cat look on his face when I first saw the Pope had written a 43,000 word thing, is it good or bad that my first reaction was to take the document and run it through Claude and say, “Can you give me a summary of what the Pope is saying about AI?” Yeah. Exactly. Yeah. There’s, yeah. Fascinating.
[Allen] You know what we don’t have yet, David? What? We need our first mortgage industry conference ran by AI. You- Robots … AI does everything. Yeah. The robots are running around, they check you in, they scan your badges and they dim the lights, they turn the music on. It’s coming.
[David] Yeah. A lot of things are coming, and it’s coming at a faster rate than I ever thought we’d see. It is encouraging, but the advantage is to those that are adopters, not just, fans of it and looking at it. You, but it’s the adoption. Which one to adopt, how fast. And I think someone said it earlier, it, the rate of adoption, or the rate of change, it’s like you can’t be afraid of making mistakes. You just gotta keep moving at it. Keep your contracts short so that you’re not getting locked up. If you can do that, you can switch around, say I don’t wanna switch.” How do you want to stay in business? You may have to in order what’s going on.
[Bill] Dave, to that point, made this comment to folks a couple of years ago and got a very blank look, but if you compare AI with autonomous driving, right? So lots of car manufacturers have done it, but Volvo is the one that sticks in my mind, Volvo has been working their customer base into getting comfortable with autonomous driving for 25 or 30 years. Yeah. Introducing things one piece at a time, and it’s “Wow, this is really neat. It helps me drive.” And it’s- Yeah … you don’t e- you don’t think about all of the individual steps that have taken place, but then you step back and go, “Yeah, it’s close.” Yeah. And I was thinking about that when, Alan’s talking about the, a pre-approval tool. That if it’s taking data and analyzing it and telling you what the customer’s information is all about, it’s not going to a true ultimate decision. Things like that are what get people over the comfort hurdle saying, “Okay, I see what you’re looking at, and I see how you’re analyzing it,” and you get more and more comfortable every day. That’s what’s gonna get it over the hump.
[David] Yep. It’s a great point. Really good point.
Allen Pollack
, Chief Operating Officer, Tech Consultant
Allen Pollack, a Mortgage & Financial Services Technology Advisor, is a subject matter expert in the mortgage origination process along with software product management and software development.
In today’s financial services push to all things Digital, Allen has been helping lenders and financial services solution providers align their digital transformation and technology strategies by removing the human element of risk, and automating processes that drive efficiencies and margins into profits.
Over the course of his career, Allen has co-created and developed technology business models that have birthed highly successful, innovative solutions and companies.
Allen co-founded and served as CTO of New York Loan Exchange (NYLX), a loan product eligibility and pricing engine (PPE) that made an immediate impact on the industry, scaling the company quickly and forming partnerships with multiple mortgage and financial lending companies. In 2012, Allen was a co-founder of a merger between NYLX and Aklero Risk Analytics that created LoanLogics, A Mortgage Loan Quality and Performance Analytics company. Allen served as CTO where he continued to bring new and innovative product solutions to the market that made a significant impact to mortgage lenders that reduced risk, scaled business channels, and grew profits in a very competitive and highly regulated market.
Allen is also is mortgage and finance technology contributor on a weekly live industry podcast, Lykken on Lending, and is launching a new podcast soon to be released, TechStack Radio, dedicated to technology and innovation in Financial Services.