Secludy raises $4M to help banks train AI without exposing customer data
What's the deal? SecludyDealroom has a profile for this one. Try Dealroom →, a San Francisco-based privacy-tech startup, has closed a $4M seed round to build a platform that lets financial services companies train generative AI models and evaluate AI vendors without exposing real customer data. Impression Ventures, a fintech-focused firm, led the round, with participation from LAUNCH, The Syndicate (Jason Calacanis), Wedbush Ventures, Precursor Ventures, Hustle Fund, Script Capital, Mana Ventures, and Chispa VC.
Why now? Every major financial services company has a mandate to ship generative AI, but the most valuable training data — transaction histories, fraud patterns, customer support logs, loan files — is locked behind privacy laws, customer contracts, and cross-border regulations. These barriers also stall vendor evaluations, which is where most enterprise GenAI deals get stuck.
Secludy sits between enterprise AI teams and the privacy, legal, and security reviews that must clear before anything ships. It lets AI teams move fast while giving compliance teams a way to say yes instead of no. Programs that took six months of review can now move in days, the company says.
"The data that would make these models most powerful is exactly the data teams can't touch," said Ben Cerchio, founder and chief executive officer. "We built Secludy to close that gap."
Cerchio previously worked in product privacy at TikTok and InfoSec compliance at PayPal. Co-founder and chief technology officer Ming He holds a PhD in computational biology with a machine learning specialisation.
What could go wrong? Secludy is entering a crowded space. Synthetic data and privacy-preserving AI tools are attracting attention from incumbents and well-funded startups alike. Winning trust in highly regulated industries like banking and insurance takes time — and a single misstep with sensitive data could undermine the company's core value proposition.
The signal: Secludy's seed round lands as privacy-enhancing technology moves from nice-to-have to infrastructure layer for regulated industries. The investor mix — spanning fintech-specialist Impression Ventures, Jason Calacanis's Syndicate, and Precursor Ventures — suggests broad conviction that the compliance bottleneck around AI training data is a durable, category-defining problem rather than a passing friction point.
Read more: globenewswire.com
