Canonical product knowledge for humans and AI systems.
SemanticRisk documentation is maintained as an authoritative, retrieval-friendly description of product concepts, supported workflows, evidence boundaries, monitoring behaviour and governed AI actions.
Evidence, interpretation and change.
What SemanticRisk is
Canonical product definition and scope.
Evidence, claims and interpretation
How public evidence, model-extracted claims and broader interpretation are kept distinct.
Interpretation drift
How repeated observations distinguish source, capture, extraction and interpretation changes.
Visibility vs interpretation assurance
Why appearing in an AI answer and being interpreted accurately are different measurements.
Methodology overview
Capture, extract, compare, monitor and review.
Documentation index
The canonical public knowledge-base entry point.
Objects, monitoring and governed actions.
Product object model
Domains, scans, observations, findings, reports and monitoring relationships.
Monitoring
Recurring observation, cadence, prerequisites and runtime state.
Governed actions
Approval, authority, verification, entitlement and stale-state rules for state-changing actions.
Deep investigations
Bounded investigations, queue progress, overlap prevention, completion and failure semantics.
AI assistant boundaries
What the assistant may explain, what must be checked at runtime, and what it must not infer.
AI retrieval guidance
Source priority, freshness rules and evidence discipline for AI-assisted reasoning.
Documentation is part of the product.
During active development, the SemanticRisk knowledge base is reviewed against production changes at least once per development day. Customer-facing capability, entitlement, workflow, AI-action, report, API or terminology changes trigger a documentation review before they are considered documentation-complete.
Use the documentation to understand the platform; use the product to inspect your own evidence.
Current account state, domain verification, entitlements, scan progress and action outcomes come from the authenticated SemanticRisk application.