← All articles
Governance

AI governance toolkit: eight practical starting points

September 2026 Muuvment
LinkedIn

Adapted from the Muuvment Labs governance toolkit for use alongside a real implementation. Copy the prompts into your working documents and assign owners. These are starting points, not completed policies, certification, or legal advice.

This web edition is a revised checklist-based adaptation, not a claim that it matches the separately circulated PDF. It removes universal scoring thresholds and fixed schedules: the controls and review frequency must fit the task.

01 · AI acceptable use policy

Define who may use which tools, for what purposes, with which information.

Fields to complete: Scope and owner; approved tools; prohibited uses; data boundaries; required review; training; incident reporting; review date.

Owner: ____________________   Review date: ____________________

02 · Project intake

Define the problem before selecting technology.

Fields to complete: Workflow; sponsor; users; baseline; intended benefit; alternatives; data requirements; acceptance criteria; budget and decision owner.

Owner: ____________________   Review date: ____________________

03 · Data readiness

Identify whether the necessary information can be used responsibly.

Fields to complete: Source owners; permissions; quality; missing information; sensitivity; retention; transfers; recovery; lineage where needed.

Owner: ____________________   Review date: ____________________

04 · Model and solution selection

Compare options against the actual task.

Fields to complete: Required controls first; representative test results; latency; integration; total operating cost; portability; staff capability.

Owner: ____________________   Review date: ____________________

05 · Risk assessment

Record risks and their owners in the specific workflow.

Fields to complete: Potential harm; likelihood; impact; existing controls; remaining risk; action; owner; approval; next review.

Owner: ____________________   Review date: ____________________

06 · Vendor assessment

Resolve supplier and data questions before commitment or sensitive-data transfer.

Fields to complete: Processing locations; access; model-training use; retention and deletion; assurance evidence; support; outages; export; contractual responsibilities.

Owner: ____________________   Review date: ____________________

07 · Ethics and release review

Check effects on people and readiness to introduce the workflow.

Fields to complete: Intended use; affected people; appropriate testing; reviewer authority; exceptions; reversal; monitoring; feedback; release decision.

Owner: ____________________   Review date: ____________________

08 · Governance responsibilities

Make decisions and escalation proportionate to the work.

Fields to complete: Decision rights; accountable owner; required technical and legal input; escalation route; review triggers; records; reporting.

Owner: ____________________   Review date: ____________________

Use the checks in the right order

Define the use case and required controls first. Resolve vendor, contractual, and data-use questions before sensitive information is transferred. Evaluate the workflow before release, and name the people responsible for monitoring, incidents, and change.

A weighted score cannot compensate for a failed mandatory requirement. Choose review dates and escalation triggers appropriate to risk and change rather than company size alone.

Start with the workflow checklist or see how these questions fit our delivery method.

NEXT STEPS

Put AI to work. Keep control of what matters.

You don’t need a fully defined AI project. Tell us where work is slow or difficult, or what you hope AI could help your business achieve. We’ll explore whether there is a useful next step.

Bring a process description, not confidential client files. Explore our resources before a conversation.