Working references
Playbooks
Practical playbooks for AI governance, security, sovereign strategy, centres of excellence, use case discovery, and workforce capability. Rules of thumb, working templates, and failure signals.
Six areas. Each page states the decision it helps you make, gives twelve rules of thumb, provides templates you can use without modification, and lists the signals that tell you the approach is failing. No theory. No case studies.
AI Governance: A Practical Playbook
Governance that produces decisions rather than principles. Decision rights before values, tiering by consequence rather than by technology, and every approval carries a stopping condition.
Templates included: AI Model Register schema, Approval Gate checklist
AI Security and Assurance: A Practical Playbook
The exposure sits in the data path and in agent authority, not usually in the model. Separate the model from the authority, and treat prompt injection as unsolved by designing so it reaches nothing.
Templates included: Vendor assurance question set, AI control checklist, incident triage sheet
Sovereign AI Strategy: A Practical Playbook
Sovereignty is a ladder of specific controls, each with a price. Name the layers you will hold, share, and accept, and budget the operating decade, not the launch.
Templates included: Sovereignty control ladder, national AI strategy test, five-year cost model
AI Centre of Excellence: A Practical Playbook
Build, enable, or govern: a function attempting all three delivers none. Staff it with people senior enough to say no, and write the sunset clause at the founding.
Templates included: CoE charter template, intake form, operating mode selector
AI Use Case Discovery, Readiness and Maturity: A Practical Playbook
Score on value, data readiness, decision ownership, and reversibility. Kill anything without a named owner who will act on the output, and cap pilots at what you can put into production.
Templates included: Four-axis scoring sheet, readiness assessment, maturity rubric
AI Leadership and Workforce Capability: A Practical Playbook
Three audiences, three curricula. Executives need practice refusing weak proposals, assessment happens by artefact rather than attendance, and internal faculty makes the capability compound.
Templates included: Three-audience curriculum map, executive session design, capability evidence framework
These playbooks sit alongside the Last-Mile AI Framework and the AI Leadership pages. The framework explains why systems fail at the point of delivery; the playbooks state what to do about it. For applying a specific standard, the interactive NIST AI RMF Playbook walks the Govern, Map, Measure, and Manage functions in depth.