Is this source actually authoritative?
Decide whether this source is the one your institution would trust in a dashboard, audit trail, policy interpretation, or official workflow decision.
Participant handout
Context readiness checklist for AI-assisted workflows.
Use this handout before connecting a document set, spreadsheet, dashboard table, policy collection, or retrieval source to an AI workflow. The goal is not to stop experimentation. The goal is to decide what is trustworthy enough to become operational context.
How to use it
Ask these questions before you trust the context.
Is it current enough to act on?
Check refresh cadence, version history, and whether someone can explain when it was last updated and what changed.
Should the system even be allowed to use it?
Confirm access rules, sensitivity constraints, privacy concerns, and whether the workflow is allowed to expose or transform this information.
Can someone defend where the answer came from?
If the source cannot be cited, traced, or explained to another colleague, it is not ready for trustworthy AI-assisted work.
Quick rubric
Classify the source before you connect it.
- Ready now: authoritative, current, permissioned, and explainable.
- Needs cleanup: useful source, but ownership, freshness, or access rules are still unclear.
- Keep out: sensitive, untrusted, or too ambiguous to support operational use safely.
What is the failure mode?
Name what goes wrong if the system uses this source uncritically: wrong answer, outdated answer, leaked answer, or overconfident answer.
Where does a person stay in the loop?
Decide whether the workflow should answer directly, cite and defer, or escalate to a staff member before action is taken.
Workshop use
A good source is not automatically a good context source.
Many institutional materials are useful for human staff but still unsafe for AI workflows because they are outdated, exception-heavy, or hard to explain. This checklist helps teams slow down just enough to decide whether a source belongs in the system, needs more governance work, or should remain outside the workflow.