He spent years inside the machinery of lending and fraud, watching the same manual work repeat. Now the Proximitty co-founder is teaching AI agents to do it - document by document, covenant by covenant.
A two-million-dollar line of credit should not take a month and a half to approve. But at most banks it still does. Tax returns get emailed back and forth. Rent rolls are re-keyed by hand. Covenants are checked against spreadsheets that someone updated last quarter, maybe. Forty-five days later, a deal that made obvious sense on day one finally clears. Wye Yew Ho spent years watching versions of this happen, and he could not stop thinking about it.
That itch became Proximitty, the company he co-founded and now runs as CEO out of San Francisco. It went through Y Combinator's Winter 2026 batch with a plain description and an ambitious one underneath it: an AI operating system for commercial loans. In practice, that means software agents that read documents, spread financials, track covenants and handle the grinding middle of lending that nobody outside a bank ever sees.
The OriginHo is Malaysian, and one line on his resume tends to make people pause: he is the highest-ranked Malaysian in the history of the Southeast Asian Maths Competition. It is the kind of credential that hints at how his mind works - patient with structure, comfortable with problems that take a while to yield.
He studied at the London School of Economics, then went into consulting at McKinsey, where his early work meant advising banks and fintechs on risk and lending strategy. That is an unglamorous corner of finance. It is also where you learn, in detail, how credit actually gets underwritten and where the process breaks. He was absorbing the shape of the problem long before he decided to solve it.
Consulting teaches you to see problems. Operating teaches you to fix them under pressure. Ho got the second education at Taptap Send, the cross-border remittance company, where he led FinCrime and Growth during a stretch when the business climbed from $75 million to $200 million in annual recurring revenue.
Those two jobs sit at opposite ends of a company. Growth pushes to sign more customers faster. Financial crime prevention pumps the brakes to keep the bad ones out. Holding both at once forces a particular discipline: you learn where automation genuinely helps and where a human judgment call is non-negotiable. That tension - speed against safety - is exactly the line Proximitty now walks in a far more regulated arena.
Consumer fintech had its decade. Payments, neobanks, buy-now-pay-later - all rebuilt for the phone. Commercial lending mostly sat it out. The reason is that the work is genuinely hard to automate. A commercial loan is full of edge cases and judgment calls that differ by loan type, borrower and covenant. Legacy software tends to choke on exactly the parts that matter.
Ho's team decided the bar could not be "close enough." Proximitty parses documents at better than 99.8% accuracy, with a human validation step designed to push that to 100 where it counts. The agents do not just read - they act. They send the borrower communications, update records and escalate the exceptions a person should actually look at.
The servicing lifecycle Proximitty's agents run end to end
The ProofThe pitch could stay abstract, except for a number the company likes to tell. One fintech client shrank its loan-servicing team from fifteen people to two. The thirteen were not cut - they were moved. Freed from document chasing, that team went and closed $20 million in new originations in a single quarter. The machine did the paperwork so people could do the deals.
One client's servicing headcount, before and after Proximitty
The adoption story has a clever twist that reflects how well Ho understands his buyers. Banks are notoriously slow to integrate anything. So Proximitty plugs into the core systems - Fiserv, FIS, Jack Henry, Finastra, Temenos and more - through browser automation rather than APIs. No IT tickets, no long procurement fight. Most implementations go live within two weeks. The bottleneck was never really the technology. It was the friction of getting in the door.
Ho did not build this alone. His co-founder and CTO is Zi Zhang, who previously led security infrastructure at Bloomberg - the systems protecting more than 300,000 terminals - and ran engineering at ACI.dev. It is a clean division of labor. Ho carries the deep knowledge of how lending and compliance actually work; Zhang carries the ability to make hardened infrastructure that a regulated bank will trust. In a category where "regulator-grade" is the whole promise, that pairing is the product.
The AmbitionThere is something contrarian about the choice. Plenty of founders chase the visible, glamorous front end of fintech. Ho went for loan servicing - the middle of the process, the part with the fewest headlines. That is often exactly where durable businesses hide, because the work is painful, the incumbents are sleepy and the switching cost, once you are in, is high.
Backed by Y Combinator, Cohen Circle, CRV and Uphonest Capital, and already working with a $5 billion publicly traded fintech and a $2 billion regional bank in its first year, Proximitty is early but moving fast. The company even turned up on a Times Square billboard during an early launch week - a loud gesture for a startup whose whole product lives in the quiet back office of banking.
Ho's stated goal is simple to say and hard to build: let lending teams spend their time on relationships instead of paperwork, and let institutions grow their loan books without growing their headcount. It is the same idea he kept circling back to at McKinsey and Taptap Send, now finally something he owns. The forty-five-day loan is the enemy. Wye Yew Ho has decided to spend this chapter of his career shrinking it, one automated document at a time.