Informational Fresh portal_guide: PRD-103

Discovery Assurance

What it is. An evidence-led walk through your declared AI estate. Where AI Discovery finds systems, Discovery Assurance assures the picture: scope, sources, signals and findings move through explicit human review, and uncertainty stays visible instead of being smoothed over.

Who it's for. Governance teams preparing a defensible statement of "this is our AI estate, and here is how we know."

Features and what they mean

Feature	What it means
Declared scope	You state intent, inclusions and exclusions up front — the assessment is honest about what it did not look at.
Sources & signals	Evidence-bearing sources are registered; declared signals are kept distinct from assessed findings, so raw input never masquerades as conclusion.
Human review queue	Findings pass role-gated human decisions — the platform records who qualified what, and with what uncertainty.
Assessment export	The reviewed state exports as an assessment pack with its provenance attached.

What you put in, what you get out

You put in	You get out
A declared scope (intent, inclusions, exclusions)	A human-governed assessment state per stage
Source references and captured signals	Findings with qualification outcomes and visible uncertainty
Reviewer decisions	An exportable, evidence-linked assessment

Your first session

Open Discovery Assurance in your workspace — the journey tiles walk the stages in order.
Declare scope — even a rough first scope makes the rest honest.
Register one source and capture one signal from it.
Walk it through Review findings and the Reviewer queue — the human-decision record is the differentiator.

Discovery Assurance workspace — Demo instance — the guided journey

Honest limits

Data isolation runs to row-level security on the database, but a fresh tenant starts honestly empty — the value builds as you feed it scope and sources.

Access: DISCOVERY_ASSURANCE entitlement; workspace at /discovery-assurance/.