Outsourcing IT doesn't outsource the obligation. See what the law expects →
Governance you can fill in
on one page.
Every template here is a job aid, not a form to submit. Print it, fill it in, and you have the beginning of a record. Keep it current on a cadence and you have a program, because governance is an ongoing program rather than a one-time audit.
A blank AI system register and the eight questions a board can ask about any AI system, with a gateway row and a reporting line.
WHY? Fill this in once and you have a record your board can read out loud. Keep it current and you have a program.
Glossary: AI system register, Proven Stop, MCP Gateway. Academy: modules M8, M20.
One governance question per AI architecture layer, the record that answers it and the accountable owner, plus a starter RACI. Editable spreadsheet and printable PDF.
WHY? Every time the blueprint gains a layer, the ledger gains a line. This is the page where you write the line, with a name and a date.
Glossary: Retrieval Provenance Record, MCP Gateway. Academy: module M20.
Who holds each AI governance responsibility today, by name: one accountable name per row, a backup and a review date, including who can stop each agent and who approves its changes.
WHY? Most organizations have the responsibilities without the titles. A regulator or an underwriter asks for the name, not the title. This page is the answer.
Glossary: Responsibility register, Agent-incident signatory. Academy: module M6.
Govern what each AI agent is permitted to do: one fill-in page per agent covering its connections, the changes it may make, its spend and who can stop it, plus the autonomy ladder.
WHY? A standard AUP governs people, not autonomous agents. Use it to give each AI agent a written job, an accountable owner, and an autonomy level before it acts on your systems.
Glossary: Agent Identity (Pillar 3), Subagent Spawning, Scheduled Agent Task. Academy: modules M13, M16.
Ten plain-English questions to ask any AI-agent vendor before you buy, plus five follow-ups for agents that connect to your systems, with good-answer and red-flag guidance.
WHY? You can't build agent controls into a vendor's product, so ask them to show they did. Use it before buying any AI-agent product to surface identity, stop and data questions early.
Glossary: Vendor Risk Management (VRM), MCP Gateway. Academy: modules M11, M13.
Ten questions for the managed provider who already holds your systems, with the same good-answer guidance and follow-ups on gateways, agents and personal accounts.
WHY? Your MSP's platform ships AI features you did not choose. Use it to find out what runs on your endpoints, who reviews it, and who is accountable for it.
Glossary: MCP Gateway, Proven Stop. Academy: module M11.
The transparency artifact an MSP hands clients before being asked, covering AI features, review, stops and escalation.
WHY? Clients are starting to ask what AI runs inside a managed service. Use it to answer that unprompted, as a trust asset.
Glossary: Inbound third-party agent, Proven Stop. Academy: module M11.
The recurring record: inventory, autonomy, off-spec events, decisions, and the stops, permissions and records checked this quarter.
WHY? A single dated report shows you looked once. Use it to show the cadence, which is what turns an assessment into a program.
Glossary: Proven Stop, AI testing evidence. Academy: modules M17, M18.
The live Academy curriculum by role, plus a half-day workshop your team can run on its own with these templates.
WHY? Governing AI agents is a new literacy. Use it to see which Academy track fits each role, and to run your team through the workshop in an afternoon.
Glossary: Agentic Zero Trust, AI literacy (EU AI Act Article 4). Academy: modules M12, M13, M16, M17.
Educational templates, not legal advice.
What a filled-in register makes possible
A regulator, an underwriter, or a director will ask for the register before asking for anything else. A register with named owners against each entry is the fastest way to answer that first question well — and the eight questions a board can answer with these artifacts show where each answer lives.