ISO/IEC 42001 certification proves that your company runs a managed, auditable AI Management System (AIMS). Getting there means scoping your AI systems, building the management-system processes the standard requires, running them for long enough to produce evidence, and passing a two-stage audit by an accredited certification body. Companies with an existing SOC 2 or ISO 27001 program move noticeably faster, because much of the operational groundwork already exists.
The steps to certification
1. Scope
Decide which AI systems, products and teams are covered, and your role for each — developer, provider or user of AI. A tight, honest scope is faster to certify and easier to maintain.
2. Gap assessment
Compare where you are today against Clauses 4–10 and the Annex A controls. The output is a prioritised remediation plan.
3. Build the AIMS
- AI policy and objectives, approved by leadership
- Roles and responsibilities for AI
- AI risk assessment and risk treatment methodology
- AI system impact assessments — effects on individuals, groups and society
- Statement of Applicability (which Annex A controls apply, and why)
- AI lifecycle processes: design, data, development, testing, deployment, monitoring
- Third-party / supplier processes for AI
4. Operate it
Run the processes for real: complete risk and impact assessments, train staff, monitor AI systems, handle issues. Certification auditors want to see the system working, not just documented.
5. Internal audit and management review
Both are mandatory before certification. The internal audit checks the AIMS against the standard; management review shows leadership is engaged and acting on results.
6. Certification audit
- Stage 1 — documentation review and readiness check
- Stage 2 — evidence that the AIMS is implemented and effective
Nonconformities found at Stage 2 must be corrected before the certificate is issued.
7. Maintain it
Certificates follow a three-year cycle with surveillance audits in the intervening years and a recertification audit at the end.
Typical timeline
| Starting point | Typical path to Stage 2 |
|---|---|
| Existing SOC 2 or ISO 27001, few AI systems | Roughly 3–6 months |
| Some security program, several AI systems | Roughly 6–9 months |
| No formal management system | 9–12 months or more |
These are planning ranges, not promises — the biggest variables are scope, how many AI systems you run, and how quickly your team can complete assessments.
What drives cost
There's no fixed price. Total cost is made up of:
- Certification body fees — driven by scope, number of sites and headcount
- Implementation effort — internal time plus any consulting support
- Tooling — often none needed beyond what you use for SOC 2
- Remediation — technical controls you don't yet have (monitoring, logging, data governance)
- Ongoing cost — surveillance audits, internal audits and keeping assessments current
The cheapest path is usually to reuse your SOC 2 controls, keep scope tight, and avoid buying new platforms until you know you need them.
How SOC 2 work speeds things up
Risk assessment, access control, change management, supplier management, incident response, logging, training and internal oversight already exist in a mature SOC 2 program. The new work is mainly AI-specific: impact assessments, AI lifecycle controls, data quality and provenance, transparency and human oversight. See SOC 2 vs ISO 42001.
FAQ
Is ISO 42001 mandatory?
No. It's voluntary. Companies pursue it because customers ask for it, or to demonstrate structured AI governance alongside regulation such as the EU AI Act.
Does ISO 42001 certification mean we comply with the EU AI Act?
No. It provides a strong management framework that supports compliance, but legal obligations under the Act must be assessed separately.
Can a small startup get certified?
Yes. The standard scales — a small company with a tight scope and a few AI systems can certify without heavy bureaucracy.
Planning an ISO 42001 path? See our ISO 42001 readiness service or book a free scoping call.