AI & Agents

Twenty-four questions across evaluation, guardrails, observability and governance. Shows your thinnest area.

A checklist and scoring sheet for reviewing AI-generated code, with 16 failure modes and what catches each.

Eighteen ways agents fail once they can act, each with a detection test and a control.

Five capabilities, five ownership gates and five metrics to commit to before you scale an AI pilot.

Put AI governance where it changes outcomes, in the delivery pipeline.

Check whether an agent is ready to act on its own, and what has to be true first.