AssurePlan Docs

Evidence and AI Boundaries

How AssurePlan keeps AI assistance grounded in plan data and reviewed evidence.

Evidence First

AssurePlan treats the imported plan, user-confirmed edits, scan findings, and evidence references as the source of truth. Reports and exports should be generated only from the current plan version and current scan state.

What AI Can Help With

AI can assist with explanation and drafting tasks, such as:

  • Summarizing why a finding matters
  • Suggesting follow-up questions for a PM
  • Drafting report language from reviewed findings
  • Explaining a what-if option and its trade-offs

What AI Must Not Do

AI must not bypass evidence gates, silently rewrite the plan, invent source-system facts, or turn stale scan output into a formal report. Customer project data is not used to train AI models by default.

Human Control

The PM remains responsible for accepting findings, confirming plan edits, selecting what-if recommendations, and exporting formal deliverables.

On this page