Compliance & audit
Frameworks & evidence
The regulations you choose become enforced behaviour, and AgentKeyRoute collects the evidence that they are.
Supported frameworks
| Area | Frameworks |
|---|---|
| Healthcare | HIPAA, HITRUST |
| Privacy | GDPR, CCPA / CPRA, India DPDP Act |
| AI | EU AI Act, ISO 42001, NIST AI RMF |
| Security | SOC 2, ISO 27001 |
| Finance | PCI DSS, SOX, GLBA, NYDFS Part 500, DORA |
| Education | FERPA |
| Government | FedRAMP, CMMC, CJIS |
From regulation to routing
Each framework carries concrete limits. HIPAA: health data only to on-prem models or providers with a BAA, records kept six years. GDPR: personal data only to processors with a DPA. EU AI Act: prohibited uses are refused and autonomous high-risk use is held for a person. Put frameworks on a profile and all of them are enforced at once; where they disagree, the strictest wins.
What an auditor will ask for
| Question | Where the answer is |
|---|---|
| Which AI is in use, by whom, for what? | Workloads (opens in a new tab): every AI use with owner, purpose, data and EU AI Act tier. |
| How is sensitive data controlled? | Profiles (opens in a new tab) and rules (opens in a new tab), with the effective policy per profile and a decision trace for any request. |
| Can you prove nothing was altered? | The audit log (opens in a new tab) is hash-chained; verify it in one click. |
| How do you track control effectiveness? | Automated checks score each framework in Frameworks & evidence (opens in a new tab); manual controls are attested with evidence references. |
| Who changed what, and was it approved? | Every configuration change is audited; two-administrator approval can be required. |
| What are your AI risks? | The risk register (opens in a new tab) scores inherent and residual risk with owners and review dates. |
Evidence pack
Under Frameworks & evidence (opens in a new tab), Evidence pack downloads framework scores, gaps, attestations and the audit-chain check as one file.