Reviewed playbook library
Start from a testable protocol, then report what happened.
These records define a method. They do not claim community outcomes. Fork one into your private workbench, run it on your context, and keep the failures as part of the evidence.
| Playbook | Track | Version | Reviewed | Evidence state |
|---|---|---|---|---|
| Run a controlled support-triage pilotCompare a human-only baseline with an AI-assisted draft-and-review workflow without automating the final customer decision. | business | 1.0 | 2026-08-24 | Reviewed protocolEditorial reference protocol; no community outcome is claimed. |
| Build portfolio proof in a weekly evidence cycleTurn one role-relevant task into a reproducible portfolio artifact with a baseline, evaluation, limitations, and verification notes. | careers | 1.0 | 2026-08-24 | Reviewed protocolEditorial reference protocol; no community outcome is claimed. |
| Design an assessment with declared AI rolesSpecify where AI is permitted, required, or prohibited, then compare learner evidence against an explicit rubric. | teaching | 1.0 | 2026-08-24 | Reviewed protocolEditorial reference protocol; no community outcome is claimed. |
| Gate a RAG change with a fixed evaluation setCompare two retrieval configurations on the same versioned cases, failure taxonomy, latency, and cost gates. | builders | 1.0 | 2026-08-24 | Reviewed protocolEditorial reference protocol; no community outcome is claimed. |