Scrutineer · Frameworks

ISO 42001 compliance software for AI management system certification

ISO 42001 is the first management system standard for artificial intelligence, and most teams already hold the evidence for a good part of it.

Scrutineer maps your existing ISO 27001 and SOC 2 controls onto the Annex A objectives, then shows you the gap that is genuinely new.

or try it below ↓

Control-mapped findings · linked evidence · you decide what to remediate

The Scrutiny Desk

Illustrative sample · not an audit attestation

SOC 2 ISO 27001 HIPAA GDPR PCI DSS

Controls in evidence-linked report out

AI scrutinizes you decide

Why it works

What you get with ISO 42001

Your ISO 27001 evidence, reused not rebuilt

ISO 42001 shares the harmonized management system structure with ISO 27001, so clauses 4 through 10 are familiar ground. Scrutineer maps the controls you already evidence onto the Annex A objectives and marks the ones your existing programme genuinely does not answer, which is a much shorter list than most gap analyses imply.

The AI-specific gap, named

Impact assessment, data provenance and life cycle verification are where an ISO 27001 programme runs out. Those objectives ask for records nobody generates by accident, so Scrutineer treats them as their own workstream with owners and dates rather than burying them in a control list.

An inventory that survives the audit

An AI management system is only as credible as its system inventory. Scrutineer keeps each AI system, the data behind it, its impact assessment and its owner on one record, so the scope you certify is the scope you can still describe twelve months later.

What it handles

Controls in, an evidence-linked report out

Point Scrutineer at a framework or a vendor and it maps every control, pulls the evidence it can find, flags the gaps and scores the risk, returning a report with linked evidence and a prioritized remediation list. Scrutineer is decision support for readiness, an accredited auditor still issues the attestation.

Maps one control set to ISO 42001, ISO 27001, SOC 2 and your other frameworks at once
Shows which Annex A objectives your existing evidence already answers, and which it does not
Keeps an inventory of AI systems with owner, data sources, impact assessment and status
Tracks impact assessments and life cycle verification records as dated, reviewable artifacts
Carries supplier and customer responsibility allocation across into the A.10 objective
Keeps the management review, internal audit and corrective action records the clauses require

ISO 42001 readiness_report

READINESS · 82%

ACCESS CONTROL 91

evidence · MFA enforced and access reviews evidenced.

CHANGE MGMT 78

evidence · Mostly covered; one approval log left untested.

VENDOR RISK 64

evidence · Two subprocessors missing a current review.

ENCRYPTION 86

evidence · Data encrypted in transit and at rest, evidenced.

Example report layout, not customer data

Why Scrutineer

One platform that maps controls and scores risk

Not a static questionnaire, not a pass-fail black box, and not a spreadsheet you maintain by hand. Live control mapping across SOC 2, ISO 27001, HIPAA, GDPR and PCI, automatic evidence and a prioritized gap list, returned as a report you can act on. The AI scrutinizes, you decide.

Mapped to real controls

Every framework is broken down into the controls it actually requires, each scored on a red to amber to green scale, so readiness stays transparent and consistent.

Evidence behind every finding

Each control links to the exact evidence that satisfies it, the policy, the config, the log line, so the finding is auditable and your readiness is defensible.

A prioritized gap list

Open gaps roll up into a ranked remediation list, so the highest-risk findings sit at the top and your team fixes what matters before the audit begins.

Annex A reference

The nine ISO 42001 Annex A objectives, and how much of each your ISO 27001 evidence already answers

Most ISO 42001 gap analyses present all nine objectives as new work. They are not. The last column is the one worth reading: it separates the objectives an existing information security programme largely covers from the three that generate records nobody produces by accident.

Annex A objective	What it asks for	Evidence an auditor samples	Does your ISO 27001 evidence already cover it?
A.2 Policies related to AI	An AI policy that is approved, communicated and reviewed, and that sits consistently alongside your other policies.	The policy itself, the approval record, the review date and evidence it reached the people it applies to.	Partly. The policy machinery transfers directly from your ISMS. The content is new, and a policy copied from a template without an AI inventory behind it reads as exactly that.
A.3 Internal organization	Defined roles and responsibilities for AI, and a route for people to report concerns about an AI system.	A responsibility assignment, job descriptions or terms of reference, and the escalation path with evidence it works.	Mostly. The governance structure carries over. The addition is naming who can decide that an AI system is fit to deploy, which is often genuinely unassigned.
A.4 Resources for AI systems	An account of the resources AI depends on: data, tooling, compute, and human competence.	A resource inventory covering each AI system, plus competence records for the people governing it.	Partly. Asset inventory and competence records exist in an ISMS, but models, training data and compute are rarely classified as assets in one, so the inventory usually needs extending.
A.5 Assessing impacts of AI systems	An assessment of the consequences an AI system can have for individuals, groups and society, not just for your organization.	A completed impact assessment per system, dated, with the decision it supported.	No. This is the single largest genuinely new requirement. A security risk assessment asks what harm can come to the organization; this asks what harm the organization can cause. ISO/IEC 42005 gives guidance on performing it.
A.6 AI system life cycle	Responsible design, development, verification, deployment and operation, with objectives set before building.	Design records, verification and validation results, release approvals and post-deployment monitoring.	Partly. Secure development controls transfer. Model verification and validation do not, and this objective carries the most individual controls of the nine.
A.7 Data for AI systems	Control over the data used to develop and run AI systems, including provenance, quality and preparation.	Data provenance records, quality checks, preparation steps and the acquisition basis for training data.	Partly. Classification and handling transfer. Provenance and quality do not, and provenance for data acquired years ago is often the hardest record to reconstruct.
A.8 Information for interested parties	Documentation and disclosure so users and affected parties understand the system and its limits.	System documentation, user-facing disclosures, and the record of what was communicated to whom.	Largely new. An ISMS documents internally; this objective is about what you tell people outside the organization, including those subject to a decision rather than buying the product.
A.9 Use of AI systems	Responsible use of AI, including how your own workforce uses AI systems you did not build.	An acceptable use position for AI, and evidence that actual use is monitored against it.	Partly. Acceptable use policies exist. Covering third-party AI tools staff adopted on their own usually does not, and shadow AI is the common finding here.
A.10 Third-party and customer relationships	Responsibility allocated clearly across suppliers, partners and customers for the AI in the chain.	Supplier agreements, the allocation of responsibilities, and supplier assessment records.	Mostly. Supplier security controls transfer well. The addition is stating who is accountable when a supplier model produces a bad outcome, which contracts written before 2023 rarely address.
Running it once for all of it	One mapped control set covering ISO 42001, ISO 27001 and SOC 2, evidenced once and reviewed on one cadence.	The same artifacts answering every row above, with the AI-specific records tracked as their own workstream.	This is where Scrutineer sits. The point is not that ISO 42001 is easy; it is that three of the nine objectives are where the real work lives, and knowing which three is worth more than another control checklist.

References are to ISO/IEC 42001:2023 Annex A, whose objectives run A.2 to A.10. Published counts of the individual controls under those objectives vary between 38 and 39 depending on how one AI life cycle sub-clause is split, so confirm the exact list against the copy of the standard you purchase. Assessments of ISO 27001 overlap in the final column are practitioner judgment about typical programmes, not a published crosswalk, and your own coverage will differ. Scrutineer prepares and maintains certification evidence; an accredited certification body performs the audit and issues any certificate.

Good questions

Questions about ISO 42001

ISO/IEC 42001:2023 is the first international standard for an artificial intelligence management system, published in December 2023. It is a management system standard in the same family as ISO 27001, so it certifies how your organization governs AI rather than testing any individual model. It combines requirement clauses 4 through 10 with a reference set of AI controls in Annex A and implementation guidance in Annex B.

It addresses governance rather than model performance. Certification looks at whether you have decided which AI systems you run, assessed their impact on people, controlled the data feeding them, verified them across the life cycle, disclosed what users need to know, and allocated responsibility with your suppliers and customers. An auditor will not tell you your model is accurate. They will test whether you can show how you decided it was fit to deploy.

Annex A groups its AI controls under nine objectives numbered A.2 to A.10: policies related to AI, internal organization, resources for AI systems, assessing impacts of AI systems, the AI system life cycle, data for AI systems, information for interested parties, use of AI systems, and third-party and customer relationships. Most published guides count 38 individual controls under those nine objectives, though counts of 39 also circulate depending on how one life cycle sub-clause is split.

No. As of August 2026 no US federal or state law requires an ISO 42001 certificate, and the state law that came closest to mandating an AI risk management programme moved away from it. Colorado delayed its AI Act to January 1, 2027 and, in SB 189 signed May 14, 2026, removed the deployer obligations to run a risk management programme and complete impact assessments. Texas enacted TRAIGA effective January 1, 2026 but pared it back to government use plus a short list of prohibited purposes. In practice US companies certify because enterprise customers ask for it in procurement, not because a regulator does.

Reported audit fees vary widely with scope and certification body. Practitioners report roughly $7,000 to $20,000 for a narrow scope, while larger certification bodies are reported to quote in the $20,000 to $50,000 range for the combined Stage 1 and Stage 2 audit. Surveillance audits are commonly reported at roughly 20 to 30 percent of the initial fee per year. None of those figures include your own preparation effort, which is usually the larger cost. Confirm pricing with the certification body directly.

Commonly reported timelines run 3 to 12 months from starting work to certificate, driven mostly by how much of a management system you already run. An organization already certified to ISO 27001 sits at the fast end because clauses 4 through 10 are already operating. An organization starting with no management system, no AI inventory and no impact assessments sits at the slow end, and the inventory is usually what takes longest.

It follows the standard management system pattern: a Stage 1 readiness review, a Stage 2 certification audit, then a three-year certificate. Surveillance audits in years one and two confirm the system is still operating, and a full recertification audit before the three years lapse starts the cycle again. The practical implication is that ISO 42001 is a recurring commitment, not a one-time project.

ISO/IEC 42006:2025, published July 7, 2025, sets the requirements a certification body must meet to audit and certify AI management systems. It supplements ISO/IEC 17021-1 with AI-specific competence requirements. It matters commercially because it is what separates an accredited certificate from a piece of paper: ask any prospective certification body which accreditation body accredits them for ISO 42001, and check that the accreditation is live rather than pending.

ISO 27001 governs information security, protecting the confidentiality, integrity and availability of information. ISO 42001 governs artificial intelligence, covering how AI systems are designed, assessed for impact, fed with data, monitored and disclosed. They share the same management system structure, so an ISO 27001 programme transfers a large share of the machinery. What does not transfer is AI impact assessment, data provenance and model life cycle verification.

These are two different things that share a name. An organization is certified against ISO 42001 by an accredited certification body. An individual takes a lead auditor or lead implementer training course and is certified by the training provider as competent to audit or implement the standard. A certified individual on your staff does not make the organization certified, and organizations buying compliance software are almost always after the first.

The clearest case is a company selling AI-enabled software to enterprises, because AI questions are now standard in security reviews and a certificate answers them once instead of per deal. It also fits organizations deploying AI in decisions that affect people, in hiring, lending, healthcare or housing, where you need to show a defensible governance process. Organizations using a handful of commercial AI tools internally usually get more value from an AI acceptable use policy and a supplier review than from certification.

Not automatically, though the work overlaps heavily. The NIST AI Risk Management Framework, released January 26, 2023, is voluntary and organizes AI risk work into four functions: govern, map, measure and manage. An ISO 42001 management system produces most of the artifacts an AI RMF programme would, which is why crosswalks between the two exist. The EU AI Act is a law with its own conformity obligations and applies on its own terms, so treat it separately if you sell into the EU.

No. Scrutineer prepares and maintains the evidence, the AI system inventory and the control mapping so the audit is short and defensible. An accredited certification body performs the Stage 1 and Stage 2 audits and issues any certificate. Anyone selling you a certificate directly from a software product is not describing how management system certification works.

Read every compliance and risk question

Keep reading

Guides that go deeper on AI governance and the standards next to it

Choosing an AI governance framework

ISO 42001 against the NIST AI RMF and the US state AI laws as they actually stand in 2026, and how to tell which one your buyers are really asking for.

Read the guide

All 93 ISO 27001 Annex A controls

The control set that carries a large share of your ISO 42001 work, listed in full with the 2022 numbering.

Read the guide

How long ISO 27001 certification takes

The certification cycle ISO 42001 mirrors, stage by stage, with the parts teams consistently underestimate.

Read the guide

Explore more

More ways to scrutinize compliance and risk with Scrutineer

SOC 2 compliance

Map controls to the Trust Services Criteria, collect evidence, and close gaps before audit.

Learn more

SOC 2 compliance software

A platform that maps SOC 2 controls, automates evidence, and tracks readiness continuously.

Learn more

ISO 27001 compliance

Map your ISMS to Annex A, automate evidence, and stay certification-ready.

Learn more

See every use case See pricing Back to Scrutineer

Stop guessing about readiness. Scrutinize on real evidence.

Point Scrutineer at a framework or a vendor and it maps every control, gathers evidence and scores the risk, returning an evidence-linked report and a prioritized gap list. The AI scrutinizes, you decide.

See pricing

SOC 2, ISO 27001, HIPAA, GDPR & PCI · evidence-linked controls · readiness, not certification