Every organisation
ISO 42001 checklist: controls, evidence and the audit trail for AI agents
ISO/IEC 42001:2023 is the international standard for an AI management system (AIMS): the policies, roles, risk and impact assessments, controls and monitoring an organisation uses to develop or use AI responsibly. Certification is voluntary and is carried out by an accredited certification body, not by ISO.
Not sure this rule covers you? See who each rule covers
The rule
ISO/IEC 42001: AI management systems, a voluntary standard
- Who it covers
- Any organisation that wants a certifiable way of managing its AI. Customers increasingly ask their suppliers about it.
- What it asks
- Run an AI management system: policies, risk and impact assessments, controls, monitoring and improvement for the AI you use or build.
- When
- When you choose to certify, or when a customer asks.
- Sources
- ISO/IEC 42001
What ISO 42001 requires
The standard follows the same structure as ISO 27001 and ISO 9001, so an organisation with one of those already has the shape of the system. The clauses ask for:
- Context: the AI you use or build, the people it affects, and your role as provider, user or both.
- Leadership: an AI policy, roles with authority, and management commitment.
- Planning: an AI risk assessment and an AI system impact assessment, with objectives and the controls chosen to meet them.
- Support: competence, awareness, documented information.
- Operation: running the controls through the AI system life cycle.
- Performance evaluation: monitoring, measurement, internal audit and management review.
- Improvement: corrective action when something goes wrong.
The Annex A control areas
Annex A lists the reference controls, grouped into areas. You choose which apply and justify the rest in a statement of applicability, as in ISO 27001:
- Policies related to AI
- Internal organisation: roles, responsibilities and reporting of concerns
- Resources for AI systems: data, tooling, computing and people
- Assessing impacts of AI systems on individuals, groups and society
- The AI system life cycle: objectives, design, verification, deployment, operation and monitoring, logging
- Data for AI systems: acquisition, quality, provenance and preparation
- Information for interested parties: documentation, incident reporting, communication
- Use of AI systems: responsible use, intended use, objectives
- Third-party and customer relationships: suppliers and customers of AI systems
An ISO 42001 implementation checklist
- Decide the scope: which AI systems, which parts of the organisation, and whether you are a provider, a user or both.
- Inventory the AI systems in scope, including the AI apps staff use with company data.
- Write the AI policy and name the roles: who owns the system, who approves new uses, who can stop one.
- Run the AI risk assessment and the AI system impact assessment for each system.
- Choose the Annex A controls, write the statement of applicability, and implement what is missing.
- Put monitoring and logging in place so that you can show each control operating, not just that it exists.
- Train the people who use and run the systems.
- Run an internal audit and a management review.
- Fix what the audit found, then book the certification audit.
Where ContextGate's record helps you evidence ISO 42001
Use this table as a starting point for your own assessment. It is not a statement of conformity, and your auditor decides what counts as evidence. ContextGate is not certified against ISO/IEC 42001.
| Area of your AI management system | What ContextGate gives you | Coverage |
|---|---|---|
| AI system inventory | The Agents page lists the AI apps and API keys using ContextGate, who signed each one in, and when it was first and last seen. | Partial. Only AI apps connected through ContextGate appear. |
| Operational controls and acceptable use | Policies on every tool call: plain-English governance checks, PII redaction, secret blocking, content-blocking patterns, and switches to turn off individual tools. | Supports tool calls routed through ContextGate. |
| Monitoring and records | The Activity log records each tool call and its policy outcome, filterable by app, connection, tool, status, policy and time, with CSV export. Attention lists the last 7 days of blocked and warned calls. | Supports. No retention guarantee and no tamper-evident log. |
| Data protection | Personal data in tool results is replaced with placeholders before the AI app sees it. Tool calls that carry secrets are blocked, and the secret is kept out of the log. | Supports tool results. Arguments are logged as the AI app sent them. |
| Roles and access | Team roles: Owner, Admin, Member and Viewer. API keys can be created and revoked. | Partial. Covers access to ContextGate only. |
| Risk and impact assessments | Not part of the product. | Not covered. |
| Management-system processes | Not part of the product. AI policy, objectives, internal audit and management review stay with you. | Not covered. |
| Supplier management | Not part of the product. ContextGate is itself a supplier you would assess. | Not covered. |
ContextGate sees only the tool calls that go through it, not an AI app's built-in tools or the chat itself. That boundary sets the limits of the evidence it can give you.
The report
ISO 42001 Evidence Report
ISO/IEC 42001, a voluntary standard
Evidence for the monitoring and control parts of an ISO/IEC 42001 AI management system. ContextGate itself is not certified against ISO/IEC 42001.
What it covers
- The activity record, as monitoring evidence.
- The policies in force and the tools switched off, as controls.
- Blocked and warned calls, as controls operating.
Who needs it
Organisations that run, or want to certify, an AI management system.
What you still do
Run the management system, and get certified if you choose to.
Questions people ask
What is ISO 42001?
ISO/IEC 42001:2023 is the international management system standard for artificial intelligence. It specifies how an organisation establishes, runs and improves an AI management system covering policy, roles, risk and impact assessment, controls, monitoring and improvement, for AI it builds or uses.
Is ISO 42001 certification mandatory?
No. It is a voluntary standard. Organisations certify because customers ask, because it structures their AI governance, or because it helps them become compliant with laws such as the EU AI Act, which it does not replace.
What is the difference between ISO 42001 and ISO 27001?
ISO 27001 is the information security management standard; ISO 42001 is the AI management standard. They share the same clause structure and can be run as one integrated system. ISO 42001 adds AI-specific requirements: impact assessments, AI system life cycle controls, data for AI, and responsible use.
What evidence does an ISO 42001 audit look for?
That each chosen control exists and operates: the AI policy, the risk and impact assessments, the statement of applicability, training records, monitoring and logs showing the controls applied to real AI activity, internal audit results and management review minutes.
Is ContextGate ISO 42001 certified?
No. ContextGate helps you gather evidence for the monitoring and control parts of your own AI management system, for the tool calls that go through it. Certification is yours to pursue with an accredited body.
A summary, not legal advice: check your obligations with your compliance team or adviser.
Ready to see what your AI agents do?
Create a free workspace and add ContextGate to your AI apps in minutes. No card required.
Get started free
We'll email you a sign-in link to set up your workspace.
Rather talk it through first? Book a 30-minute call