Notes on AI agent governance, evidence, and what enterprise security teams should be asking in 2026.
A security survey, two AI-lab frameworks, and two governments arrived at the same four questions this summer. Almost none of today’s tooling can answer all four.
Most enterprises in 2026 are buying Shadow AI protection for a problem that has two layers — and getting coverage for only one of them.
Most AI governance products rely on the agent to cooperate with its own oversight. In 2026, that assumption is starting to break in public.
Most products in the AI governance category solve the wrong problem — and the difference will matter more in 2026 than it did last year.