As enterprises troubleshoot their AI systems, governance has emerged as a key challenge. Some are taking a dual approach: one model to handle incoming queries, and another to keep the first one from getting into trouble.

That’s the premise of ZeroDrift, a new AI compliance service that on Tuesday said it had raised $10 million in a seed funding round that saw investments from a16z Speedrun, Reign Ventures, Pitchdrive, and U&I Ventures, among others. The company deals entirely with the second part of the system, sitting between AI models and end users to flag and replace any messages that might present a compliance problem.

It might seem strange to build an AI tool to correct other AI systems’ mistakes, but ZeroDrift says its system has a few architectural advantages over the models it will be correcting. The system is triggered by conventional programs that deterministically apply known compliance standards like SOC 2 or GDPR, and the LLM only comes into play once a message has been flagged, rewriting a compliant version of the same message.