Enterprise AI Is Entering an Evaluation Gap: Agents Are Gaining Autonomy Faster Than Companies Can Verify Them
VentureBeat, Friday, July 10th, 2026
AI agents are gaining production autonomy faster than enterprises can build trustworthy verification for them.
A June 2026 VB Pulse survey of 157 enterprise respondents found half had deployed an AI agent or LLM feature that passed internal evaluations yet still caused a customer-facing failure, with one in four failing more than once.
Meanwhile 66% already allow, or plan within 12 months, production deployment without human review, but only 5% fully trust the automated evaluations behind those release decisions.
This mismatch is the 'evaluation gap': the autonomy ceiling is rising faster than the assurance beneath it.
Passing a test does not guarantee real-world correctness, as agents can pull correct data but update the wrong fields. The piece urges better verification methods before accelerating agent deployment.