ISO 42001 readiness.
We get your AI management system ready for an ISO/IEC 42001 audit: the scope, the risk method, the controls and the evidence. An accredited certification body does the certifying, not us.
Readiness, not certification
Who certifies, and who doesn’t.
Mono is not a certification body, an auditor or a law firm. Not sure you need the standard at all? Read whether your company needs ISO 42001, and how it compares with SOC 2.
ISO
Publishes the standard. ISO does not issue certificates to companies.
Certification bodies
Independent firms that audit you and decide whether to certify. ISO/IEC 42006:2025 sets requirements for the bodies that audit and certify against 42001.
Accreditation
Accreditation bodies check the certification bodies. In India that is NABCB. Elsewhere it is bodies such as ANAB or UKAS.
Consultants, including Mono
We provide readiness. We can’t issue a certificate, and we can’t make a certification body’s decision for it.
What readiness covers, and what we build.
We do the engineering and the evidence, so the audit tests real systems. The same records also help with AI security questionnaires, such as the AI sections of the SIG and the CSA’s AI-CAIQ. See how those questionnaires map to AI controls.
Scope
What we checkWhich AI systems, teams, vendors and locations the management system covers.
What we buildA written scope statement the auditor can test against.
AI policy and roles
What we checkWho owns AI risk today, and what rules already exist.
What we buildA short AI policy, named owners, and a review rhythm that actually runs.
AI risk and impact assessment
What we checkHow you judge risks and effects on people for each AI feature.
What we buildA repeatable method, applied to your real features and recorded.
Controls
What we checkWhich of the standard’s Annex A controls you already meet, in practice and on paper.
What we buildEngineering work for the gaps: logging, approvals, data handling, vendor settings.
Evidence
What we checkWhether each control leaves a record an auditor can open.
What we buildEvidence collected from your systems, filed by control, easy to hand over.
Internal audit and review
What we checkWhether anyone has checked the system yet.
What we buildA dry run of the audit, so the first outside audit holds no surprises.
A Statement of Applicability, worked through.
Clause 6.1.3 of the standard asks for a Statement of Applicability. It lists the controls you need, and gives a reason for each one you include or leave out. This is an illustration for a made-up SaaS feature that summarises customer support tickets using a provider’s model. It shows the shape, not a finished document, and your own risk assessment decides the real answers. Annex A has 38 controls, and not every one is mandatory.
A.2 Policies related to AI: Include
WhyAn AI policy exists to be read and followed.
EvidenceApproved policy, version history, staff acknowledgement.
A.3 Internal organization: Include
WhySomeone must own AI risk and report on it.
EvidenceRole descriptions, review minutes.
A.5 Assessing impacts of AI systems: Include
WhyThe feature summarises customer text, so errors affect customers.
EvidenceImpact assessment for the feature, signed off.
A.6 AI system life cycle: Include
WhyChanges to prompts, models and settings need control.
EvidenceChange records, test results, release approvals.
A.7 Data for AI systems: Include
WhyCustomer data reaches the model provider.
EvidenceData flow diagram, provider settings, retention configuration.
A.10 Third-party and customer relationships: Include
WhyThe model comes from an outside provider.
EvidenceProvider terms, due diligence record, customer-facing documentation.
Any control that only applies to in-house model training: Exclude, if true
WhyOnly valid if you call a provider’s model and train nothing yourself. The reason is written down.
EvidenceArchitecture note confirming no in-house training.
How it runs.
These are phases, not promises of dates. How long each takes depends on your size, your AI features, what already exists, and the certification body’s calendar. Readiness often sits inside wider AI governance work.
- Step 1
Gap check
We compare what you have with the standard, and agree the scope with you.
- Step 2
Build
We write the policy and risk method, and do the engineering for the controls you lack.
- Step 3
Evidence
We collect proof from your systems and complete the Statement of Applicability.
- Step 4
Dry run
We run an internal review and fix what it finds.
- Step 5
Handover
You choose an accredited certification body. We support you through their audit questions.
Related services.
AI security questionnaire help
Answer the AI sections of enterprise questionnaires, using the same evidence.
See questionnaire helpShadow AI audit
Find AI tools used without approval, so your scope matches what really happens.
See the shadow AI auditAI governance
The full set of AI governance work for SaaS companies selling to enterprises.
See AI governanceISO 42001 FAQs.
Who issues an ISO 42001 certificate?
An independent certification body, after an audit. ISO publishes the standard but does not certify companies, and neither do we. Only bodies accredited for ISO/IEC 42001 issue accredited certificates, for example by NABCB in India. We prepare you and hand over the evidence.
Is ISO 42001 required by law?
It is a voluntary standard. Whether you need it depends on what your buyers and regulators ask for. If your customers have not asked, a smaller set of controls and good questionnaire answers may be enough first. Whether a contract or sector rule requires it is a question for your counsel.
What is a Statement of Applicability?
A document that lists the controls you have chosen from the standard, says why each is included or left out, and points to the evidence. The standard asks for one as part of AI risk treatment. Auditors read it early, so it needs to match what you really do.
How long will readiness take?
It depends on your size, how many AI features you run, what controls already exist, and when your chosen certification body can audit. We don’t promise a timeline up front. After the gap check we can tell you what the work involves.
Do we need ISO 42001 if we already have SOC 2?
They answer different questions. A SOC 2 report is an opinion from a licensed CPA firm on your controls. ISO 42001 is a management system standard for AI, and its certificates come from accredited certification bodies. Some buyers ask for one, some for both, and some for neither.
Is this legal advice?
No. We work on the engineering and the evidence: systems, settings, logs and documents. Contract terms and legal questions go to your counsel, or to our partner law firm if you want us to involve them.