Enterprise IT Agents Miss More Than Half — The Empty Seat Called the Verification Layer

The Problem

Enterprises are told to hand IT operations to autonomous agents, yet frontier models get more than half of SRE incident-resolution tasks wrong. Worse, more investigation turns produce more false positives, not better answers.

Why Now

The fact that agents fail is itself the market. There's an empty seat for a verification, human-in-the-loop, and scope-narrowing layer that lets teams run agents safely at 'propose' authority instead of 'execute.'

Recommended Talent

Engineers who've carried an SRE on-call pager, paired with someone who has built eval pipelines for grading LLM-agent output.