Pavan Agarwal’s Approach to Fair Lending

The mortgage industry has spent years producing statements about commitment to fair lending. Pavan Agarwal has spent years producing data about it, and the gap between the two is where his companies have made a specific argument about what that phrase actually requires.

Sun West Mortgage Company, the national lender Agarwal leads, operates under what it calls a no-additional-overlays policy. Overlays are supplementary requirements that individual lenders layer on top of agency lending guidelines, including additional credit score floors, stricter debt-to-income thresholds, and documentation requirements that go beyond what is actually mandated. They are legal. They are widespread.

And, as Agarwal has argued publicly, they disproportionately affect borrowers who are already less likely to have textbook-perfect credit files: first-generation homebuyers, non-traditional income earners, and members of communities that have historically faced barriers in the lending process.

By eliminating overlays, Sun West evaluates every applicant against the same standards the agencies themselves set. The result, according to the company, is a decline rate for minority borrowers that is approximately half the national average.

The same philosophy carries through to the technology Agarwal has built.

AngelAi®, the AI platform Sun West runs through Celligence LLC, applies consistent evaluation criteria to every borrower, removing the variability that can enter human underwriting and the unconscious bias that variability can carry. The platform operates in over 100 languages, and handles the documentation burden in natural conversation rather than requiring borrowers to navigate industry vocabulary on their own. Sun West fully warrants every underwriting decision the system produces, which means the lending institution itself stands behind the AI’s output rather than disclaiming responsibility for it.

AngelAi was also built to be aware of how users feel during the process, not only what they’re asking. The system detects nervousness, reassures users when it senses stress, and escalates to a human executive when it senses frustration. Celligence has trademarked the concept as Empathetic Technology. The design reflects Agarwal’s view that fair lending is partly about how borrowers experience the process, not only whether they are approved at the end of it.

The funding model carries the same logic. AngelAi is sustained through the wholesale lending it enables, not through fees charged to the people who use it. Agarwal applies a test to every product decision that captures the standard he holds the platform to: can a grandmother use it? He means it literally. The system was designed for people with no prior experience navigating financial software and no advisors to walk them through. People in exactly that position have completed the entire home financing process through AngelAi.

Agarwal frames the work consistently as a mission rather than a market position. He grew up in an immigrant family in California that ran a real estate business, and he watched the mortgage process operate as a barrier to first-generation buyers.

His father Hari Agarwal had come to the United States from India in 1973 and started over in his forties. Pavan and his sisters worked in the family business from childhood, attending Cerritos public schools and helping at the office after class. He has said that the gap between the aspiration to own a home and the process of obtaining one has always struck him as a problem with a technical solution. AngelAi is his attempt at that solution.

The platform has now processed more than 210,000 real transactions.

The fair lending conversation in the mortgage industry has stayed at the level of corporate statements for years. Agarwal has tried to move it into operational practice, where the policy choices, the product design, and the lender’s accountability for outcomes can all be examined together.