I grew up in Queens after immigrating to the United States from India at the age of eight. Queens is a place where hundreds of languages are spoken, where people arrive from all over the world, learn a new language, adopt a new culture, and work relentlessly to build a better life. Watching that happen up close shaped how I think about opportunity, capitalism, and the responsibility that comes with building systems that either unlock progress or quietly prevent it. You see very early that success is not just about effort or talent; it is about access. About whether the systems people rely on are designed to help them move forward or force them to fight uphill at every step.
That perspective is why I care so deeply about mortgage lending. There are many ways to make money in finance, but very few allow you to directly help families build stability, dignity, and long-term security. Homeownership remains one of the most powerful mechanisms for doing exactly that. When the system works, it brings people into the American dream and gives them a tangible stake in their future. When it doesn’t, frustration builds, trust erodes, and the distance between those who can navigate the system and those who cannot grows wider. Better was born out of that belief, and out of personal frustration with how broken the mortgage process felt when I experienced it firsthand.
Before founding Better, I spent my career managing and valuing billions of dollars of mortgage-backed securities. I understood where the data lived, how loans were underwritten, how risk was assessed, and how quickly decisions could theoretically be made. Then I tried to get a mortgage pre-approval from a major bank and was told it would take three weeks. That disconnect was jarring. Credit checks existed. Income verification existed. Automated valuations existed. Pricing grids existed. The delay had nothing to do with risk. It was bureaucracy layered on top of bureaucracy, duplication masquerading as diligence, and systems that were never designed to talk to each other. Machines were not talking to machines, so people were checking other people’s work, printing disclosures, re-entering the same data, and stretching a process that could take minutes into one that took weeks.
The original idea behind Better was not radical. It was almost obvious: If software could value, bundle, and trade mortgages in real time in capital markets, it should certainly be able to help a consumer get approved quickly, transparently, and confidently. That insight launched a decade-long effort to rebuild mortgage lending from the ground up, starting not with new products or marketing, but with the underlying mechanics of how decisions get made.
The first phase of that journey was digitization. We took paper, phone calls, and manual workflows and put them into a single digital flow. That was the work of roughly 2015 through 2020, and while it was necessary, it was not sufficient. Digitizing a broken process does not fix it; it just makes the brokenness faster. The next phase was automation. Rules-based systems replaced thousands of repetitive human tasks. Roughly seventy percent of consumer-facing actions became automatically triggered and resolved by software. This was never about eliminating people. It was about letting humans focus on work that actually required judgment instead of acting as glue between disconnected systems.
The phase we are in now is what I would call AI-native lending, and that distinction matters. In most organizations, humans do the work and AI assists them. At Better, the AI does the work first. When it fails, a human steps in. Humans are the exception, not the default. That only works because every task across our platform has historically been performed inside the system. We were not training models on hypothetical scenarios or clean academic datasets. We trained them on real underwriting decisions, real exceptions, and real judgment calls made by experienced professionals over many years. The AI learned not from theory, but from watching how humans actually behaved when the rules ran out.
None of this means humans are becoming irrelevant. If anything, it clarifies where humans matter most. There are parts of the mortgage experience that machines should not replace. Empathy. Reassurance. The ability to understand a family’s hopes, fears, and tradeoffs. Many of our retail loan officers begin relationships with what they call a hopes-and-dreams conversation. They talk about why someone wants to buy, what success looks like to them, and how homeownership fits into the rest of their life. AI does not do that. People do. What AI does is remove friction behind the scenes. It turns hours of analysis, spreadsheets, and back-and-forth into instant clarity. It allows loan officers to spend time connecting rather than compiling. That division of labor is the future: tasks best done by machines should be done by machines, and tasks best done by humans should be protected and elevated.
Where the industry still struggles is its obsession with complexity. More products. More acronyms. More channels. More internal specialization. All of this can feel sophisticated inside an organization, but it is deeply misaligned with how consumers think. Borrowers do not care whether a loan is FHA, VA, conforming, or non-QM. They care about two numbers: how much money they need upfront and what the monthly payment will be. Everything else is noise. For years, Better made its own mistakes here. We were arrogant. We believed consumers should transact one way, through one channel, on our terms. That was wrong. What we learned from working with retail loan officers and partner lenders is that people want to be met where they are. Some want digital. Some want in-person. Many want both. The future is not forcing behavior. It is adapting to it.
What most consumers actually want is not a faster mortgage. It is a better one. They want private-bank-level service at a mass-market price. They want to feel respected, understood, and guided. They want technology to work quietly in the background, the way travel or delivery apps do. They do not want to learn how the system works. They want the system to work for them. That is why the industry’s fixation on labels misses the point. Digital mortgage. AI mortgage. Broker mortgage. Direct-to-consumer mortgage. None of that matters to borrowers. Value matters. Simplicity matters. Trust matters.
Looking ahead, buying a home will start with a conversation, not an application. A consumer will describe what they want, what matters to them, and what they can afford. Software will surface options, explain trade offs, and eventually ask the most important question in complex transactions: do you want me to do this for you? When confusion arises, a human will step in. When reassurance is needed, a human will step in. But most of the mechanical work will fade into the background. Much of what exists today in mortgage technology stacks will not exist a decade from now. That is not a threat. It is an opportunity.
I do not believe success comes from being right the first time. It comes from refusing to make the same mistake twice. Better today is not about winning at the expense of the industry. It is about helping the industry win. Most of our growth now comes from empowering other lenders and fintechs through our AI platform. When our partners succeed, we succeed. We are willing to self-disrupt, abandon old assumptions, and meet customers and partners where they actually are, not where we think they should be. That mindset is not just a strategy. It is a necessity.
Mortgage lending is not about transactions. It is about relationships that begin at closing, not end there. It is about helping people manage the most important asset they will ever own. Technology should reduce friction, not replace humanity. If we get that balance right, we do more than modernize an industry. We help keep the American dream alive for the next generation.