For Legal

GenZAgents for compliance officers — evidence packs that auditors actually accept

Auditors increasingly ask about AI activity in your org. The evidence packs auto-generated from GenZAgents receipts satisfy SOC 2 + ISO 42001 + EU AI Act + EU CRA in one bundle — and they're signed so the chain of custody is verifiable without contacting us.

The "no AI audit trail" finding

Most orgs failing AI audits do so on one specific finding: "no audit trail for AI-assisted work". The auditor can't verify which AI was used, who supervised it, what it did. Without a credible answer, the finding stands. Remediation: either a 6-month engineering project to build internal AI logging or a GenZAgents deployment that ships in a week.

What the evidence pack contains

Per-control receipt mapping. Signature verification proofs. Agent registry with DIDs and KYC levels. Anomaly history + responses. Dispute history + outcomes. Per-provider data-flow documentation. The bundle is a signed zip your auditor can verify offline (Ed25519 + JCS, standard verifiers available).

Multi-framework in one pack

SOC 2 (CC6.1, CC7.1, CC7.2, CC9.2, 2026 AI addendum). ISO 42001 (Clauses 6-10, Annex A). EU AI Act (§50 transparency, §10 data governance, §13 record-keeping). EU CRA (Article 13, Annex I §3.1). Same evidence base; different control mapping. The pack generator handles all four out of the box; you customise via /admin/compliance.

Period-of-time vs point-in-time

SOC 2 Type 2 wants 6 or 12 months of period-of-time evidence — receipts cover that natively. ISO 42001 wants point-in-time process documentation — the pack assembles it from agent registry + anomaly history + dispute resolution data. One pack covers both audit styles.

Auditor acceptance

In our experience: most auditors accept the signed receipt-based evidence pack with no friction. The chain of custody is cryptographic; the verification is standard tooling. The exception: auditors unfamiliar with the receipt format may want a 30-minute walkthrough — we provide a 1-page "how to verify a GenZAgents evidence pack" doc that gets them oriented.

Compliance officer's 5-minute gut check

Next audit cycle: are you in scope for SOC 2 / ISO 42001 / EU AI Act / EU CRA AI controls? If yes, your audit prep effort is significant. GenZAgents Enterprise (£6k/year) cuts the AI-evidence-gathering effort from weeks to days. The ROI is in the audit-prep time saved, not just the audit pass.

Common questions

Do we still need a security audit on GenZAgents itself?

Yes — same vendor-risk review as any vendor handling your data. See /security for our attestations (SOC 2 Type 1 complete, Type 2 in progress, pen test annual). Standard DPA available.

How does this compare to OneTrust / Drata / Vanta?

Different scope. OneTrust / Drata / Vanta automate the broader compliance program (policies, vendor reviews, control evidence). GenZAgents specifically covers the AI-side evidence. We integrate with Drata / Vanta via webhooks to feed receipt-based evidence into their unified control dashboards.

Can the evidence pack be customised per-auditor?

Yes — the YAML control mapping is editable. If your specific auditor wants specific evidence formatted specifically, customise once and the pack uses your template.

What about FedRAMP?

Not in scope today. FedRAMP requires US gov-cloud deployment; we're hosted on Azure UK South. If you need FedRAMP-compliant deployment talk to us about on-prem (Enterprise tier).

Related

Get the trust layer for your AI work

GenZAgents is the verified work-history layer above every AI provider your team uses. Sign cryptographic receipts, hand off conversations across Claude / ChatGPT / Cursor / Gemini, keep institutional AI knowledge when employees leave.

Last reviewed · 3 min read· Open spec· Changelog