From hype to help: How Tavant is applying AI where mortgage friction actually lives

In this conversation with HousingWire’s Allison LaForgia, Dr. Olav Laudy, Tavant’s head of AI, discussed his path into mortgage technology and how Tavant differentiates its approach to AI in an industry saturated with hype.

“I am a traditionally machine learning guy,” Laudy said. “Then, naturally, you get into AI right when machine learning transferred into the AI domain. I joined Tavant, and Tavant has a mortgage arm. And naturally I was curious on how the things were already automated and how I could help automate that more using AI.”

Landy’s curiosity quickly turned into a practical philosophy around implementation. “We deploy a lot of actual mortgage professionals, people with business backgrounds in mortgage, and they understand what the friction points are,” he said. “So when I propose something, they can say yay or nay to it. And that allows me to develop in the right direction.”

While discussing strategy, Tavant does not deploy AI for the sake of novelty; their approach is more intentional.“We’re never deploying AI for the sake of AI,” Laudy said. “You need to deploy AI because it solves a friction point. Anything that takes a lot of time for someone can be potentially enhanced by an AI feature.”

One clear example of a friction point is the communication of loan status between underwriters and loan officers. “A simple email like that will take, if done properly, 20 minutes,” Laudy said. “So we built this AI agent called ‘chat with loan.’ You say, ‘write a note for the loan officer,’ and the AI reviews all the elements in the loan — what is missing, what’s the status, what’s the time to close — and writes that email.”

Even as technology accelerates the process, human oversight and control remain. “The underwriter looks through that email and says, ‘this is right’ or ‘this is not right,’” Laudy said. “Once done, they click send. That process now takes three minutes instead of 20 minutes, but it’s still a complete, relevant update.”

Additionally, that same philosophy extends to the borrower experience through Tavant’s mortgage avatar, Maya. “Maya basically accompanies you from the start of your loan all the way till the end and into the servicing end of it,” Laudy said. “She can help you fill out forms, answer questions along the way, and do important checks.”

If a borrower uploads the wrong document, the system responds instantly. “It is immediately noticed and fed back to you,” Laudy said. “Compare that to the traditional way, where it hops back and forth and takes two or three days. The borrower lives in anxiety thinking, ‘will I get my loan or not?’”

Maya also helps preempt underwriting conditions. “If an avatar scans your bank transactions and says, ‘explain this now because I know questions are coming,’ you just save valuable time for both the borrower as well as the underwriter,” Laudy said.

On compliance, Laudy stressed that AI is a tool, and not a cause for alarm. “If you would replace the word AI with computer, you realize how people are freaking out over compliance,” he said. “It is not the computer who decides who gets a loan or not. You always have a human in the loop as a final approver.”

Looking ahead, Laudy is optimistic about “policy as code” for Tavant.  He explains, “You have the borrower’s details and your guidelines, and you have an AI that very smartly goes through this.” And highlights the current strength of AI, noting, “You’d be surprised how strong AIs are today to take very complex rules and apply conditions to see if things work out.”

As Laudy summarizes Tavant’s goal:  “The goal is smoother, faster, more concise compliance — not handing everything off to a big machine and hoping for the best.