01 / Start with a useful outcome
Choose a specific team and a repeatable task, such as finding approved guidance or preparing a project summary. Define what a useful answer looks like before introducing new tools.
- Name the audience, task, and content sources.
- Record the current time and effort involved.
- Agree how people will review AI-assisted work.
02 / Understand the information estate
Map the SharePoint sites, Teams workspaces, and documents that support the task. Look for duplicate guidance, unclear ownership, and material that is no longer current.
- Assign an accountable owner to each priority source.
- Separate approved guidance from working drafts.
- Make titles and metadata meaningful to the people using them.
03 / Review access before expanding discovery
AI readiness includes information protection. Review existing access and sharing settings with the people responsible for security and the business content. Do not assume that enabling an AI tool fixes an existing permissions problem.
- Check broad access groups and external sharing.
- Review sensitive or high-impact content first.
- Document remediation decisions and who will action them.
04 / Run a measured pilot
Use representative questions and real work, with people who understand the source material. Capture incorrect, incomplete, and unexpectedly useful results, then use those findings to improve both the content and the working practices.
- Agree success measures and a review date.
- Train participants to check source material.
- Track issues with named owners rather than relying on informal feedback.
05 / Make readiness an ongoing practice
Keep content ownership, review, access, and lifecycle decisions visible after the pilot. Expand when the evidence supports it, rather than treating deployment as the end of the work.
- Review the pilot with business and technical stakeholders.
- Prioritise the next information improvements.
- Plan adoption support and recurring governance reviews.