Variance Raises $21.5M Series A for AI Investigative Agents in Compliance

Every system looks sharp until pressure hits. Then the cracks start talking. Alerts stack, signals blur, and suddenly “efficient” turns into expensive theater. Variance, formerly Intrinsic, just pulled in $21.5M in Series A funding out of San Francisco, with Ten Eleven Ventures leading the round and 645 Ventures, Y Combinator, Urban Innovation Fund, and Okta Ventures doubling down. Total raised now sits around $26M, which tells you this is not a science project. This is infrastructure with teeth.

Karine Mellata and Michael Lin are not guessing their way through fraud and compliance. They built inside Apple’s fraud and algorithmic risk machine, where mistakes are expensive and scale is unforgiving. Now Karine Mellata as CEO and Michael Lin as CTO are taking that same mindset and aiming it directly at financial crime, KYC, KYB, AML, and everything that keeps compliance teams up at night.

Variance does not sell dashboards that look pretty in board meetings. It sends in agents. These AI investigative agents do the work, pulling data from over 150 sources, tracing beneficial ownership, checking sanctions, reading messy documents, and assembling evidence like a prosecutor who does not sleep. The result is not a score. It is a case file. Audit ready, regulator friendly, and built to hold up when someone asks uncomfortable questions.

The platform claims it can gather about 90% of the evidence per case and cut investigation cycles by up to 10×. That is the difference between a team drowning in alerts and a system that actually closes the loop. Fortune 500 companies and financial institutions are already leaning in, not because it sounds good, but because it clears queues that used to take weeks.

There is a bigger lesson buried in this raise. The winners in AI are not just generating answers, they are taking accountability. Variance is shifting the game from probability to proof. From “this looks risky” to “here is exactly why.”

Capital will go toward deepening the infrastructure and expanding into more financial institutions, which makes sense. Fraud is not slowing down. Regulation is not getting lighter. And AI generated abuse is just getting warmed up.