The Agentic Readiness Gate: A diagnostic of your organization's readiness to deploy and govern agentic AI, scored on evidence, one use case at a time

The Agentic Readiness Gate is a paper diagnostic that scores how ready an organization is to run one agentic AI use case, in any industry. It asks eight readiness checks about governing the system, not the model's output, whether one agent or several do the job. Six control areas cover risk appetite and enforcement, workflow design and recovery, the final human check, data provenance, agent identity and entitlements, and evaluation and monitoring. Two value areas cover business outcome and unit economics. Each area is scored 1 to 5 against written descriptions at levels 1, 3 and 5: tick the level 3 statements you can show, none 1, some 2, all 3; all of level 3 plus some of level 5 is a 4. The ceiling is the lowest control score, never an average, following the non-compensatory logic of CMMI, ISO/IEC 33020 and release gates such as batch release. A value area below 3 means do not scale, even if the ceiling allows it; the ceiling says what not to do yet and never certifies a deployment as safe. Unknowns count as 1, marked unverified. Each score must point to evidence. The output is a ceiling and two actions, each with an owner, a date and the evidence that shows it is done. This record contains the scorecard, a method note, a fictional worked case (an analytics request agent at two companies), and an action helper. The instrument is a directional self-assessment, not a validated audit. Version 2.6 is its first public release. It will first be used with life sciences data teams at the AnalytiCon 2026 pre-conference workshop in Boston on 13 October 2026; the worked case comes from that setting.