AI Governance Boundary

What ANCHOR is

ANCHOR is governance, trust, learning, intelligence, and readiness infrastructure for safe AI use in veterinary clinics.

• Metadata-only by default.
• Human-review based: AI-supported outputs remain subject to professional judgement before operational use.
• Receipt-backed: governed activity is explainable through governance metadata.
• Trust-surface oriented and multi-tenant, with clinic-scoped separation.
• Standalone now, and architected for vendor-neutrality / vendor-neutral over time.

What ANCHOR is not

ANCHOR is not a clinical product and must not be used as one. In particular, ANCHOR is not:

• diagnostic AI
• prescribing AI
• treatment-planning AI
• autonomous triage
• an ambient scribe
• an EHR or PMS
• a clinical decision-support product
• a GPAI (general-purpose AI) model provider
• a compliance guarantee
• a regulator-approved product
• a replacement for veterinary judgement

What governance receipts are

Governance receipts are governance evidence only. They record metadata about governed activity, for example identifiers, timestamps, policy version, and review state, so that AI use can be reviewed and accounted for.

What governance receipts do not prove

A governance receipt does not prove clinical correctness, patient safety, professional competence, or regulatory compliance. It evidences that a governance process took place; it does not validate the clinical content or outcome of any underlying work.

Generation posture

ANCHOR provides deterministic governed generation today. Live Workspace generation remains production-off pending a local/staging safety gate. See the AI Data Use page for how AI providers and data are handled.