AI Management System Certification
ISO 42001 consulting for companies that build or use AI and now have to show a customer, an auditor or a regulator how that AI is governed, with particular depth in regulated industries. Gap assessment, implementation support and certification audit preparation.
"Keep it Simple. Keep it Real."
Who This Is For
ISO/IEC 42001 is the certifiable standard for an artificial intelligence management system. The companies asking about it tend to fall into two groups: those that build AI into what they sell and are being asked to prove it is governed, and those that use AI inside work a regulator or auditor already watches. We work with both, and we do our best work where the two meet.
Already have ISO 27001 and want to add ISO 42001? It is a smaller project than starting from nothing, because ISO 27001 already gives you the management-system clauses, a risk process and a Statement of Applicability. The new work is the AI system inventory, AI impact assessments, the Annex A controls that apply and records of human oversight. Certify Consulting works alongside your existing ISO 27001 or SOC 2 program, starting with a gap assessment (two to three weeks, usually $3,500 to $10,000), with implementation scoped from the gap report; four to twelve months to certificate is typical, and companies with a working management system are usually at the short end.
An ISO 9001 certified manufacturer starting to use AI: ISO 42001 is voluntary, but you have a head start because it uses the same Annex SL structure as ISO 9001, so most of clauses 4 through 10 already exist in your system. What you add is the AI system inventory, AI impact assessments, the Annex A controls that apply and evidence of human oversight, and certification is issued by an accredited body after a Stage 1 and Stage 2 audit. Certify Consulting, led by principal consultant Jared Clark, starts with a gap assessment (two to three weeks, usually $3,500 to $10,000), then scopes implementation from the gap report. Four to twelve months from gap assessment to certificate is typical, and companies with a working ISO 9001 system are usually at the short end.
In April 2026 FDA issued its first warning letter citing inappropriate use of artificial intelligence under 21 CFR 211.22(c): specifications, procedures and master records written by AI and released without qualified review. ISO 9001 and ISO 13485 auditors are asking the same question in their own words. If AI touches your records, someone will ask who approved them. See ISO 42001 for GMP manufacturers.
Enterprise buyers started writing ISO 42001 into vendor security reviews and questionnaires in 2026. Sometimes they want a certificate; more often they want documented alignment with the standard's controls. If your product includes AI, the answer has to exist before the questionnaire arrives, and your team has to be able to defend it on a call. See ISO 42001 for AI and software companies.
ISO 42001 is built on the same Annex SL structure as the standards you already certify to. Most of clauses 4 through 10 exist in your system today. What is missing is the AI-specific content, and adding it is a much smaller project than starting a management system from nothing. See ISO 42001 for ISO 9001 certified companies.
What The Standard Is
ISO/IEC 42001:2023 was published in December 2023. It describes an AI management system: the policies, roles, risk assessments, records and reviews an organization uses to keep the AI it develops or uses under control. Clauses 4 through 10 follow the same pattern as ISO 9001, ISO 27001 and ISO 13485, which is why companies holding those certificates have a head start. Annex A adds the AI-specific controls, grouped into nine areas that cover policy, roles, resources, impact assessment, the AI system lifecycle, data, information for interested parties, use of AI systems and third-party relationships.
Certification is issued by an accredited certification body after a Stage 1 and Stage 2 audit, exactly as with ISO 9001. Fewer than a hundred organizations held a certificate in January 2026 and roughly 350 did by spring, so the auditors are still few and the queues are real. The standard is not required by any US law. It also does not, on its own, satisfy the EU AI Act: the European version was adopted in March 2026 but has not been cited in the Official Journal, and the standard being written to carry the AI Act's quality-management requirements is prEN 18286. We explain that in ISO 42001 and the EU AI Act.
In our view the useful way to think about ISO 42001 is not as an AI project at all. It is a quality-system question applied to a new kind of tool: what does it do, who owns it, what could go wrong, how do we know, and what is the record. Companies that already answer those questions for their equipment and suppliers can answer them for AI. That is the whole of our approach. Read What is ISO 42001? for the longer version.
How We Work
Two starting points and one flat-fee project. You know the price before we begin, and you own everything we produce.
Quoted up front
$3,500–$10,000
A two-to-three-week review of your AI systems against every clause and every Annex A control, ending in a written report, a prioritized plan and an honest estimate of the time and cost to certification.
What the gap assessment coversFixed fee
$5,000
For GMP, ISO 9001 and ISO 13485 sites. We find where AI is already writing, checking or deciding inside your quality system and tell you whether those records would survive an inspection.
What the review coversScoped after the assessment
Quoted
Policies, the AI system inventory, impact assessments, the Statement of Applicability, training, internal audit and a mock certification audit. Priced from the gap report, so you pay for the gaps you actually have.
How implementation runsThe Process
The same eight-step method we use for every management system, with the AI-specific work in the places the standard puts it.
Inventory every AI system in use, including the ones nobody told IT about, and measure your current practice against each clause and control.
An AI policy with an owner, the objectives behind it, and the documented information clause 7.5 expects, built inside your existing document control.
AI risk assessment and AI impact assessments for the systems that affect people, products or decisions, using the risk method your system already has.
The Statement of Applicability, the Annex A controls that apply to you, supplier AI terms, and approval workflows for AI-assisted work.
Awareness for everyone who touches AI and competence for the people who own it, recorded in the training system you already run.
Human oversight where it matters, change control for AI systems and prompts, monitoring, and a record trail an auditor can follow.
A full internal audit against the standard and a mock Stage 2 so the certification audit holds no surprises.
The measures clause 9 asks for, a management review that actually reviews AI, and the improvement loop that keeps the certificate.
Why Us
Most firms offering ISO 42001 come from information security. They know ISO 27001 and they know software companies. Our practice comes from the other side: quality systems and regulatory affairs for FDA-regulated manufacturers, where the questions ISO 42001 asks about AI are questions we have been answering about equipment, software and suppliers for years. For AI and software companies that means working alongside the security team and the ISO 27001 or SOC 2 program you already have, and adding the management-system discipline those programs do not fully cover.
Jared Clark, our principal consultant, is a PECB Certified ISO/IEC 42001 Lead Implementer. He also holds a JD and an MBA, is a Certified Manager of Quality/Organizational Excellence, Certified Quality Auditor and Certified Pharmaceutical GMP Professional through ASQ, holds the PMP, and is certified in Computer System Validation, Computer Software Assurance and Pharmaceutical Validation Management. That last group matters here. Deciding whether an AI tool is fit for use inside a regulated process is a validation question before it is a governance question, and the answer has to satisfy 21 CFR Part 11 and GAMP 5 as well as ISO 42001.
He has also written two white papers on the subject, both published on Zenodo in 2025: The Process Protocol Layer (PPL), on AI that is transparent and auditable enough for regulated work, and Artificial Intelligence Exchange (AIX), a protocol for how AI agents hand work to one another.
Credentials and verification links are on the About page.
Guides
The questions every buyer asks, answered in enough detail to plan a budget and a timeline.
Included At No Additional Cost
Every ISO 42001 engagement includes access to our LeanISO platform for the AI system inventory, impact assessments and risk register, document control, internal audit scheduling and corrective actions. You keep it after certification.
with every engagement
No. ISO 42001 is a voluntary international standard. Demand for it comes from customers and procurement teams asking vendors for it, from insurers, and from regulators asking how AI inside a regulated process is controlled. It does not by itself satisfy the EU AI Act; the standard being written for that purpose is prEN 18286.
Four to twelve months from gap assessment to certificate is typical. Organizations with a working ISO 9001, ISO 13485 or ISO 27001 system are usually at the short end because the management-system clauses already exist. Certification bodies have reported auditor backlogs, so booking the audit early matters.
Our gap assessment usually costs $3,500 to $10,000, and our AI-in-the-quality-system review is a fixed $5,000. The whole project is quoted as one flat fee, usually $10,000 to $40,000 for a small or mid-sized company. Certification body fees are separate and depend on the body and the size of the audit.
A great deal. ISO 42001 uses the same Annex SL structure, so context, leadership, planning, support, operation, performance evaluation and improvement already exist in your system. What is new is the AI-specific content: an inventory of AI systems, AI impact assessments, Annex A controls and evidence of human oversight.
Yes. The standard covers organizations that develop, provide or use AI systems. A manufacturer using a vendor's AI in quality, planning or documentation is within scope, and the controls on supplier AI and on reviewing AI-generated records are the ones that matter most for that company.
Smaller than starting from nothing, and the overlap differs. ISO 27001 already gives you the management-system clauses, a risk process and a Statement of Applicability, so ISO 42001 mostly adds AI-specific content. SOC 2 gives you control evidence and audit habits but not the management-system clauses, so there is more to build. Either way the new work is the AI system inventory, AI impact assessments, the Annex A controls that apply, and records of how models and prompts are changed and overseen.
In April 2026 FDA issued a warning letter citing inappropriate use of artificial intelligence under 21 CFR 211.22(c): AI-written specifications, procedures and master records released without qualified review. For a GMP manufacturer, governing AI is now an inspection question, and ISO 42001 is the management-system way to answer it.
Schedule a free 30-minute consultation. We will ask what AI you use, what systems you already certify to, and who is asking you for ISO 42001, then tell you plainly what certification would take.
Or email us at support@certify.consulting
If you already hold ISO 27001, adding ISO 42001 is an extension of the system you run today, not a second one built from scratch. It usually takes roughly four to twelve months to certification, at the short end with a working 27001 system and a narrow AI scope. ISO 42001 is the certifiable standard for managing AI: the policies, roles, risk checks and records that keep the AI you build or use under control.
Clauses 4 through 10 carry over, and the new work is the AI system inventory, the AI system impact assessments and the 38 Annex A AI controls. Both standards share one management-system structure: context, leadership, planning, support, operation, performance evaluation and improvement. Your document control, internal audits, management review and corrective action all carry over. Not everything about those clauses transfers, though, because your risk method now has to cover harm to the people an AI system affects, and not only harm to information.
The first new piece is the AI system inventory: a list of every AI system you build or use, including third-party tools and AI inside your own products, with an owner, a purpose, the data it touches and the vendor. A spreadsheet is enough. In my experience it usually turns up tools that staff adopted without IT ever hearing about them. An auditor will expect it, and everything else builds on it.
The second is the AI system impact assessment, which ISO 27001 has no equivalent for. It asks what an AI system could do to the people affected by it, such as a job applicant screened by a hiring tool, and what you do about it. For each system that matters, it's a short written document with the owner, who is affected, what could go wrong and what limits the harm.
The third is Annex A, the list of AI controls in ISO 42001. It has 38 controls in nine areas, covering policy, roles, resources, impact assessment, the AI system lifecycle, data, information for interested parties, meaning what you tell customers and regulators about the AI, use of AI, and third parties. The lifecycle area means controlling how an AI system is designed, tested, changed, monitored and retired. Your 93 ISO 27001 controls don't substitute for these, although your security controls do help with some of them. You write a new Statement of Applicability, which lists each control and says whether it applies, and why.
The standard sets no duration, and published estimates vary. I'd plan on roughly four to twelve months to a certificate, toward a year if you have many AI systems and unclear ownership. Most of the effort is your own team's time building and running the AI controls before the auditor arrives. Your 27001 certificate doesn't count toward 42001, so an accredited certification body still runs a Stage 1 and a Stage 2 audit. Few auditors are qualified for 42001 so far, so ask your certification body now about its 42001 availability and Stage 1 dates, and build that wait into your plan.
As for who does both, most firms offering ISO 42001 come from information security, so they know 27001 well. Ask any of them whether they've written AI impact assessments, mapped the Annex A controls, and taken a client through a 42001 certification audit. At Certify Consulting we come from quality systems and regulated manufacturing, and we work alongside your security team and your 27001 program. Jared Clark, who leads it, is a PECB Certified ISO/IEC 42001 Lead Implementer.
We start with a fixed-fee gap assessment at $9,750. It takes two to three weeks, and you get a written report, a prioritized plan, and an estimate of the time and cost to certification. Implementation is then quoted from that gap report. The link to book a free call is in the description.
To start today, write down every AI tool your company builds or uses, with an owner next to each one. Then bring that list to your 27001 internal audit lead, or to a call with us.