Ask for the facts first
Request system lists, AI tools in use, data categories, policies, owners, incidents, vendor records, and existing deployment notes.
Communication kit
Client evidence communication kit
Plain-language patterns for requesting evidence, explaining review work, and following up without hype.
Explain the discovery session
Frame discovery as a factual review of current environment, workflow, and disclosure posture before any recommendation is made.
Set expectations for stakeholders
Tell interviewees that the goal is to understand real work, known gaps, and evidence sources, not to sell a predetermined AI project.
Prepare public-answer wording
Help the organisation state where AI is used, what humans review, what data is in scope, and what remains unknown.
ReadoutSummarise findings without overclaiming
Separate verified facts, partial evidence, open questions, and recommended next evidence steps.
Send the evidence gap list
Close the loop with owners, missing documents, unresolved claims, review dates, and the route where public evidence will live.
Tone
Use language that can stand in the public record.
- Say what was inspected and what was not inspected.
- Name the source of a claim when you can; mark it as unresolved when you cannot.
- Avoid confidence theater: do not call a system safe, compliant, or ready unless evidence supports the claim.
- Explain controls in business language: who reviews, who approves, what is logged, and what happens when something fails.
- Point readers to public resources, disclosures, or readiness records instead of a commercial conversion path.
Related routes
Use the kit with the evidence resources.
Discovery
Microsoft 365 discovery
Explain the data collected by the discovery scripts and how it will be used in a readiness record.
AssessmentAssessment facilitation guide
Use the guide to structure stakeholder conversations and evidence requests.
ProofEvidence proof preview
Check whether the final communication points to inspectable proof rather than broad claims.