AI Meets the Mortgage Maze: Inside Pavan Agarwal’s Bid to Modernize Lending with Real-World Data

On a rainy Monday morning in Midtown, the hum of financial offices is punctuated by something new: artificial intelligence parsing mortgage applications at speeds Wall Street banks could only dream of a decade ago. At the center of this technological leap is Pavan Agarwal, CEO of Celligence LLC, whose AngelAi® platform is quietly recalibrating how lenders assess risk and compliance.

Agarwal’s team didn’t simply code up a few algorithms and call it a day. Over the last two years, engineers and data scientists at Celligence fed AngelAi® hundreds of thousands of anonymized loan files—each one a labyrinth of figures, supporting documents, and personal histories. Layered atop this data trove: a thicket of lending regulations, spanning not just federal statutes but the granular, ever-changing requirements from Fannie Mae to the New York State Department of Financial Services.

This granular approach matters in New York, where the city’s real estate market is a world unto itself. “A Brooklyn brownstone application looks nothing like a Manhattan co-op deal or a Queens two-family home,” Agarwal explained in a recent sit-down at Celligence’s Flatiron coworking hub. AngelAi®’s edge, he says, is its ability to quickly identify red flags and regulatory tripwires specific to New York’s patchwork of housing stock.

The stakes are high: September’s uptick in mortgage applications—spurred by slightly lower rates and a wave of autumn listings—has lenders scrambling to process paperwork without missing compliance deadlines. For borrowers, especially first-timers handling the city’s labyrinthine housing market, faster and more accurate decisions could mean the difference between locking in a dream home or losing out in a bidding war.

As the city’s financial sector leans into digital transformation this fall, Agarwal’s experiment with real-world data and regulatory nuance is being closely watched—not only by banks and brokers, but by regulators and consumer advocates eager to see if AI can make lending fairer as well as faster.