Early in his career, Will Lawrence worked on the product team at Facebook focused on anti-money laundering. The former PayPal president David Marcus had been brought in to help kickstart the payments initiatives company-wide, including through WhatsApp launches in India and Brazil. But Lawrence quickly learned that the main roadblock was the decidedly unsexy compliance work of making sure the product was adhering to local “know your customer” provisions and fraud prevention.
After working on the compliance team at the stablecoin infrastructure company Paxos, Lawrence decided to ride the generative AI wave and enter one of the first Y Combinator batches after the launch of ChatGPT. His thesis was that anti-money laundering and know-your-customer compliance operations would be one of the breakout use cases for applying AI to financial services. Lawrence’s bet turned out to be prescient. Less than three years later, his startup Bretton AI (previously called Greenlite) is announcing its $75 million Series B funding round led by Sapphire Ventures, with participation from his seed and Series A backer Greylock, along with Thomson Reuters Ventures and Canvas Ventures.
Lawrence says that the world of financial monitoring has two layers. The first is risk detection, which can be solved with more rudimentary machine learning. In other words, if a user starts sending $50 million a day, a system should easily pick up that it requires further investigation. The second layer, risk remediation, is trickier. That’s where the complex investigation takes place to figure out the background of the parties involved in suspicious transactions and whether they violate a company’s internal risk policies—and where Bretton is focused.
As AI increasingly commoditizes software (and potentially makes untold numbers of SaaS startups and public companies obsolete), Sapphire managing director Rajeev Dham said that the sensitivity of a product like Bretton, which needs to sell into the trust infrastructure of massive financial institutions, “feels more protected.” Seth Rosenberg, general partner at Greylock, added that rather than using Anthropic to develop their own applications, a bank like JPMorgan could benefit from having a third-party. “When compliance businesses reach scale, sometimes they can get smarter because they’re seeing data across the entire industry,” he said. For now, Bretton has focused on financial services companies as its customers, ranging from fintech startups like Mercury, Ramp, and Robinhood, as well as a number of community and regional banks including the tech-friendly Lead Bank.
Financial services have long been a thorny arena for vertical AI companies. An AI system going wonky in software development or design might lead to a bug. The stakes are higher in finance, but the difficulty is part of the mission.
“The easy thing is to go sell marketing AI,” Lawrence joked. “The hard thing is to solve things that really matter.”
Leo Schwartz
X: @leomschwartz
Email: leo.schwartz@fortune.com
Submit a deal for the Term Sheet newsletter here.
Joey Abrams curated the deals section of today’s newsletter. Subscribe here.
