The name change alone tells you something about ambition. What started as Greenlite AI—modest, functional—is now Bretton AI, a nod to the 1944 conference that reshaped the global financial system. And with $75 million in fresh Series B funding announced Monday, the San Francisco startup is betting that artificial intelligence can do something similar for the unglamorous world of financial crime compliance.
More than 1.2 million compliance investigations now run through Bretton AI's platform at banks overseen by the Office of the Comptroller of the Currency, the FDIC, and the Federal Reserve. That's a remarkable footprint for a company that launched just over two years ago, in 2023, and it raises a question the financial services industry has been circling for months: Can AI agents actually handle the tedious, high-stakes work of spotting money laundering and suspicious transactions?
Bretton AI's early customers—Robinhood, Mercury, Gusto, and a handful of regional banks—seem to think so. One FDIC-regulated institution claims it slashed loan origination timelines by as much as 90 percent. Another, with more than $15 billion in assets, says it cut $5.35 million in business process outsourcing costs. A Fortune 500 client halved the time it takes to onboard new institutional relationships.
Those are the kinds of numbers that get venture capitalists interested. Sapphire Ventures led the Series B, with partner Rajeev Dham taking a board seat. Existing backers Greylock, Y Combinator, Thomson Reuters Ventures, and Canvas Ventures joined in, alongside newcomer TIAA Ventures. Total capital raised now sits at roughly $95 million—a steep climb from the $4.8 million seed round the company pulled together in November 2023, back when it was still called Greenlite AI.
The Series B comes just nine months after Bretton AI closed a $15 million Series A led by Greylock in May 2025. That's a tight fundraising cycle, even by the standards of a bull market for enterprise AI companies.
Automating the Tedious, High-Risk Stuff
What Bretton AI is selling, in essence, is software that can do the work of junior and mid-level compliance analysts. Its agents—autonomous, not just assistive—automate what the industry calls Level 1 and Level 2 investigations: the initial triage of alerts for anti-money laundering violations, know-your-customer checks, sanctions screening, and ongoing transaction monitoring.
The agents operate inside tools banks already use, producing audit-ready documentation meant to satisfy regulatory guidance from the OCC, the Federal Reserve, and the New York Department of Financial Services. That last part is crucial. Regulators don't take kindly to black-box AI making decisions about whether to flag a transaction as suspicious, so Bretton AI has built what it calls "Trust Infrastructure"—a governance layer designed to embed regulatory standards directly into the agents' workflows.
CEO Will Lawrence, who built compliance systems at Paxos and worked on anti-money laundering and payments at Meta, co-founded the company with Alex Jin, an engineer who came from Dropbox. Lawrence describes financial crime as "the breakout use case for AI in financial services," though he's careful to frame the technology as something that needs "the right trust and governance foundations" to work in production at heavily regulated institutions.
Translation: We know regulators are watching.
Bretton AI says it has saved customers more than $10 million in compliance-related costs and eliminated 195,000 manual hours. Those figures are hard to verify independently, but they track with what other enterprise AI startups are claiming as they push into back-office operations. Compliance work, by some estimates, consumes 10 to 15 percent of institutional budgets—a massive line item that has only grown since the 2008 financial crisis ushered in a wave of post-crisis regulations.
The Rebrand and the Bigger Bet
The shift from Greenlite AI to Bretton AI isn't purely cosmetic. It signals a pivot toward something more expansive: establishing a standard for how AI should be deployed, governed, and audited across the financial system. Dham, the Sapphire partner who led the round, put it bluntly: Bretton AI is "defining the trust infrastructure" for AI in banking.
That's a bold claim, and one that will require buy-in not just from banks but from regulators who remain cautious about handing critical compliance functions over to algorithms. The OCC has issued guidance on model risk management; the Fed has warned about over-reliance on automated systems; NYDFS has been particularly vocal about the need for human oversight in transaction monitoring.
Bretton AI's argument, essentially, is that its agents are transparent and auditable in ways that traditional machine learning models are not. Whether regulators—and examiners who show up for annual audits—will agree remains an open question.
The company, which went through Y Combinator's Summer 2023 batch, has integrated with platforms like Alloy and Middesk to streamline business due diligence and know-your-business workflows. It's hiring enterprise account executives in San Francisco and New York, a sign that it's moving upmarket to larger, more complex institutions. Those are the banks with the most to save—and the most to lose if something goes wrong.
What Comes Next
Bretton AI plans to use the new capital to expand beyond its current suite of compliance tasks, deepen relationships with regulators, and push deeper into larger institutions. That last part is where the real test will be. Regional banks and fintechs are one thing; money-center banks with sprawling compliance operations and decades of legacy systems are another.
Lawrence says the company is focused on "accelerating adoption" and continuing product development, which in startup-speak usually means scaling the engineering and go-to-market teams. The company hasn't disclosed revenue, though with a $95 million war chest and a growing roster of enterprise clients, it's reasonable to assume it's past the early product-market fit stage.
Perhaps the timing is right. Compliance work has long been viewed as a cost center—necessary but not strategic. If AI agents can genuinely handle the routine investigations that tie up thousands of hours of analyst time, banks will find the ROI compelling. The question is whether they'll trust the agents enough to let them run unsupervised. And whether regulators will let them.
For now, Bretton AI is betting they will.
