01 — REDACTION
PHI redaction first
All 18 Safe Harbor identifiers scanned before any call. Unredactable PHI blocks it.
AI security
No model call without PHI redaction, governance metadata, and qualified provider review. Baseline, not a premium tier.
Responsible AI posture
01 — REDACTION
All 18 Safe Harbor identifiers scanned before any call. Unredactable PHI blocks it.
02 — REVIEW GATE
No AI output is final without qualified provider approval.
03 — AUDIT LOG
Prompt hash, output hash, model ID, redaction status — append-only.
04 — PROVENANCE
Provider, model ID, request ID, classification — recorded per call.
05 — INJECTION DEFENSE
Hard block on token-boundary, jailbreak, review-bypass patterns.
06 — RETENTION TAGS
Tags map AI outputs to your data-handling policy.
Governance metadata
Recorded per call
Available to reviewers
Enterprise-ready AI
ChartNav is engineered around enterprise-grade AI infrastructure that security and IT teams expect — and that ophthalmology practices can actually deploy.
IBM WATSONX
Enterprise-grade AI infrastructure designed for healthcare workloads.
GOVERNANCE
Provider, model, and request metadata recorded per call.
CLOUD READY
AI, storage, and governance patterns aligned with IBM Cloud deployment.
FHIR R4
FHIR R4 read-through for interoperability with existing EHR systems.
SBOM
SBOM published per release for procurement and vulnerability review.
ROADMAP
Enterprise SSO (SAML / OIDC) and SCIM provisioning on the roadmap.