> ## Documentation Index > Fetch the complete documentation index at: https://standard.aiuc-1.com/llms.txt > Use this file to discover all available pages before exploring further. # Introduce and scope AIUC-1 > Identifying the right client, introducing AIUC-1, assessing readiness, and scoping and pricing the engagement. export const InlineClock = ({city, timeZone}) => { const [time, setTime] = useState('--:--:--'); useEffect(() => { const update = () => setTime(new Intl.DateTimeFormat('en-US', { hour: '2-digit', minute: '2-digit', second: '2-digit', hour12: false, timeZone }).format(new Date())); update(); const id = setInterval(update, 1000); return () => clearInterval(id); }, [timeZone]); return

{city} {time}

; }; **The job: find the agent worth certifying, and the reason its owner moves now.** Most clients have several AI agents. Only some are worth a certification, and the difference is how much damage it can do. | Filter | What to look for | | ------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Blast radius | The agent is critical to the business and brand is on the line when it fails | | Capability | The agent holds sensitive data, or acts through tools like a CRM or payment processing | | Obligation | The agent is subject to legal, compliance, or stakeholder requirements. An AI recruitment system classified as high-risk under the EU AI Act may require third-party validation before deployment | Two accelerants worth checking before the first call. An agent already inside an ISO 42001 scope carries roughly 20-30% overlap with AIUC-1 evidence. That shortens evidence collection materially, and it means the client has already decided AI assurance is worth paying for. Certification converts into shorter security and procurement cycles on enterprise deals, plus a co-marketing moment. That's the outcome the client is buying, so that's what to open on. Next: book the introduction meeting. **The job: make the client see a gap they can't close with what they already have.** Open on the gap. Every client already holds frameworks that govern how AI systems are built and how people are meant to interact with them. None of them look at what the agent actually does once it's live. | Risk area | Existing frameworks | AIUC-1 | | ----------------------------------- | ------------------------------------------------- | ------- | | How AI systems behave in production | None | Covered | | How AI systems are built | NIST AI RMF, OWASP Top Ten, MITRE ATLAS, CSA AICM | Covered | | How humans interact with AI systems | ISO 42001, EU AI Act | Covered | AIUC-1 is comprehensive, auditable, technically grounded, and updated quarterly. It is shaped quarter by quarter through the [AIUC-1 Consortium](https://www.aiuc-1.com/consortium), which unites CISOs, GRC leaders, practitioners, and academics from leading organizations around the world. When a client asks who decided any of this, that page is the answer. The standard covers six principles, each broken into the sub-risks the evaluations target. | Principle | Sub-risks | | -------------- | ------------------------------------------------------------------------------------------------------------- | | Data & privacy | IP infringement, personal data leakage, business data leakage | | Security | System information disclosure, unauthorized actions, insecure outputs, insecure actions | | Safety | Physical harm, emotional harm, inappropriate content, high-risk domain engagement, vulnerable user engagement | | Reliability | Service disruption, hallucination, incorrect tool calls | | Accountability | Transparency | | Society | Cyber | Three things certification surfaced for companies that have already been through it: * A customer service agent company saw its hallucination rate drop from 11% to \<2% by strengthening its groundedness filter * A customer support agent company saw inappropriate tone and output format outputs reduce from 9% to \<2% by strengthening defensive prompting and configuring output moderation * A product onboarding agent company discovered and patched a PII exposure vulnerability during the certification process **When the client says they already have ISO 42001.** Agree with them, then show the overlap. AIUC-1 incorporates the majority of ISO 42001 controls, so their certificate is an accelerant rather than a competing spend. | Principle | Where ISO 42001 already covers it | What AIUC-1 adds | | -------------- | ------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------ | | Data & privacy | Annex A.7 data management, A.4.3-4.5 resource documentation | Specificity on data preparation methods | | Security | Clauses 8.1-8.3 operational planning, A.6.2.4 verification and validation | Pre-deployment testing and third-party adversarial robustness assessment | | Safety | Clauses 6.1-6.3 and 8.2-8.4 risk assessment and treatment, A.5 impact assessment | Risk taxonomy and monitoring | | Reliability | A.6.2.4 verification and validation | Hallucination and tool-call testing via third-party evaluation | | Accountability | Clauses 4-10 management system, A.6.2.6-2.8 documentation and logging, A.8.4-8.5 incident communication | The deepest overlap of the six | | Society | A.5.5 societal impact assessment | Scoped to cyber and catastrophic misuse prevention | A yes to all of these means the agent is inside the box: certification proceeds on standard pricing and the standard timeline. Anything else is still certifiable, it just costs more time. | Question | Effect on scope | | ------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Is the agent text-first, covering inputs, outputs, and tool calls? | Text and input documents are standard. Generating output documents adds eval complexity. Voice, video, and image expand the applicable controls and require additional eval harnesses | | Can the agent be accessed via API, from external infrastructure, with a persistent API key and clear documentation? | Determines eval volume and timeline. UI-only or locked-down environments are still certifiable, but slower | | Is there a production or equivalent test environment the client can configure and seed with synthetic data? | Evals must be representative of real usage | | What are the highest rate and concurrency limits to the model providers and the agent? | Higher limits allow greater volume and more complex evals in a shorter window | | Can the client's team collect evidence independently, with engineering support and a scope approver? | AIUC publishes typical evidence descriptions for every control, so clients can collect evidence themselves | Send this when the client wants clause-level detail. Live. Send this to the client's technical owner before the scoping session. Live. **The job: turn interest into a scoped opportunity.** An outside-in gap assessment scores a company against the AIUC-1 requirements using only evidence the company has published, plus anything handed over directly. It separates requirements that already have public evidence from those plausibly covered but unproven, and from those that still need to be verified. The output earns its keep three ways. It gives the client a concrete reason for the next meeting, it names the gaps that justify the engagement, and it shows the delivery lift before anything is priced. Coming soon **The job: get sign off on scope, timeline, and cost. Then book a delivery slot.** **Agree the dates first.** Agree what happens, when, what they receive, and who owns it. That's what a client needs before they can approve a number. Kickoff to certificate takes about ten weeks. | Phase | Timing | What happens | Deliverable | Owner | | ---------------------------- | ------------- | --------------------------------------------------------------------------- | ------------------------------------------------------------- | ------------------------- | | Pre-work | Now | Confirm certification scope, and map AIUC-1 to what the client knows today | Scope and readiness assessment | Client | | Audit readiness | Weeks 1-2 | Close gaps on requirements marked to be verified, and set up sandbox access | Controls implemented: an audit-ready agent and infrastructure | AIUC | | Eval and evidence collection | Weeks 3-7 | White-box control testing and black-box adversarial testing | Eval results, every finding graded by severity | Auditor and AIUC | | Results review | Weeks 7-8 | Review deliverables and the evals dashboard, then Q\&A on findings | Remediation plan covering what to fix and what can wait | Client, auditor, and AIUC | | Certification | Week 9 onward | Audit report issued, quarterly re-testing begins | AIUC-1 certification | AIUC | One round of evals. A second round runs only if P0 or P1 findings are still open after remediation, and adds roughly three weeks. **Then price it.** One proposal, scoped and priced separately: the agent in scope, the certification fee, and the audit fee. **Then book.** Delivery capacity is reserved once the client signs. Booking the audit and certification is what moves the engagement to [Deliver and certify](/auditors/deliver-and-certify). Coming soon Coming soon Client-facing assets sit in the [Client kit](/auditors/client-kit).

37.782274° N -122.392147° W FIG. A (SITE INDEX)

Artificial Intelligence Underwriting Company

Code Structural unit a. AIUC-1 requirements for agent data, privacy, security, safety, reliability, accountability, and societal risk. b. Evidence templates for technical implementation, legal policy, operational practice, and third-party evaluation. c. Crosswalks to AI regulations, standards, and security frameworks. d. Quarterly updates shaped by enterprise adoption, risk, regulation, and community input.

I. Standard

II. Learn

III. Office