The EU AI Act is Now Applying. Does Your Business Need ISO 42001 Certification?

A team discussing AI governance around a laptop in a meeting

Key Takeaways

  • The EU AI Act does not require every business to obtain ISO 42001 certification. Obligations depend on the organisation’s role and how it develops, supplies or uses AI.
  • ISO 42001 gives organisations a structured way to manage AI risks and responsibilities, but certification alone does not prove compliance with the Act.
  • Review your AI activities and customer requirements to decide whether certification would strengthen your governance, support procurement or help build trust.

What is the EU AI Act About?

The EU AI Act sets rules for artificial intelligence across the European Union, with the aim of protecting people’s health, safety and fundamental rights. It takes a risk-based approach: it bans certain uses of AI, sets stricter requirements for high-risk systems and requires transparency in some situations, such as when people interact with a chatbot.

The rules also depend on an organisation’s role, such as developing or using an AI system. ISO/IEC 42001 is a separate, voluntary standard for managing AI across an organisation. Certification to it is not a blanket requirement under the Act.

Does The EU AI Act Apply to UK Businesses?

Yes. The EU AI Act can apply to a UK business even if it has no office in the EU. It covers providers that place AI systems or general-purpose AI models on the EU market. It can also cover providers and deployers based outside the EU when their system’s output is used in the Union. The duties depend on the organisation’s role and the system’s use and risk category.

For a UK technology company selling into Europe, checking scope early can help avoid surprises around market access and customer requirements. The GDPR has already made territorial scope a practical business issue for UK firms. The AI Act has its own test, so assess each AI system and your role under the Act.

At Tempo Audits, we audit AI management systems against ISO 42001. Our certification audit assesses the management system; it does not determine whether the Act applies to your business or provide implementation consulting.

When Does the EU AI Act Reach Your UK Business?

The EU AI Act can apply to UK businesses through several routes. Under Article 2, check whether:

  • You provide an AI system or general-purpose AI model for the EU market, or put it into service there.
  • Your organisation deploys an AI system from an establishment or location in the EU.
  • You provide or deploy an AI system from outside the EU, and its output is used in the Union. A UK SaaS company could fall within this rule when an EU customer uses an AI feature, even if the company has no EU office or staff.
  • You import or distribute AI systems in the EU.
  • You manufacture a product with an AI system and place it on the EU market under your own name or trade mark.

The duties depend on your role and the system involved. An EU customer alone does not automatically settle the question; you need to check whether the feature is an AI system and how it is used.

Two colleagues at an office desk reviewing work on a computer

Which Role Does Your Business Play Under the EU AI Act?

The EU AI Act assigns duties according to the role a business plays in relation to an AI system. A UK company may hold more than one role, depending on what it develops, sells or uses.

Provider

A provider develops an AI system, or has one developed, and puts it on the market or into service under its own name; a UK SaaS company selling its own AI feature to EU customers may be a provider.

Deployer

A deployer uses an AI system under its authority; a UK recruitment firm using AI to screen candidates for hiring decisions in France may be a deployer.

Importer

An importer established in the EU places an AI system from a non-EU provider on the market; an Irish subsidiary bringing its UK parent company’s branded AI system into Ireland may be the importer.

Distributor

A distributor in the supply chain makes an AI system available on the EU market; a French reseller supplying a third-party AI tool to local businesses may be a distributor.

Product manufacturer

A manufacturer that places a product containing an AI system on the EU market under its own name may have duties as a product manufacturer; a UK company selling AI-enabled machinery in Germany under its brand is one example.

A company’s role attaches to its activities and systems, so it may be a provider for one tool and a deployer for another.

EU AI Act Risk Levels: What The Four Tiers Mean

The EU AI Act groups AI systems by the level of risk they pose. The category depends on the system’s purpose and use, not simply on the technology or the company that built it.

Unacceptable risk

The Act prohibits certain uses that threaten people’s rights or safety. Examples include social scoring and some forms of harmful manipulation. Breaches of these prohibitions can attract fines of up to €35 million or 7% of the offender’s total worldwide annual turnover, whichever is higher.

High risk

High-risk systems include some AI used in recruitment, education, essential services and critical infrastructure. Organisations can use these systems, but providers must meet applicable requirements for risk management, data governance, documentation, record-keeping, human oversight and system performance.

Conformity assessment and database registration depend on the system and its route to market. Under the current timetable, rules for Annex III systems apply from 2 December 2027; those for high-risk AI in certain regulated products apply from 2 August 2028.

Limited risk

Some AI uses carry transparency duties. For example, people generally need to know when they are interacting with a chatbot. The Act also sets disclosure or labelling requirements for certain AI-generated content, including deepfakes. These rules apply to specified uses, not every AI output.

Minimal or no risk

The Act adds few requirements specifically because a system falls into this category. Other rules may still apply based on the system, the organisation’s role or the context in which it is used.

General-purpose AI models also have their own obligations under the Act. They sit alongside the four risk tiers rather than forming a fifth tier.

EU AI Act Compliance Timeline: What Applies Now and What Comes Next

The EU AI Act has been taking effect in stages. As of October 2026, its general application has begun, but some high-risk AI requirements have later deadlines. The Digital Omnibus entered into force on 27 July 2026 and set dates for those rules.

DateWhat it means
2 February 2025Prohibitions on certain AI practices took effect, and the AI literacy provision began applying.
2 August 2025Governance rules and obligations for general-purpose AI models began applying.
2 August 2026The Act became generally applicable, and enforcement began. Some requirements have separate deadlines.
2 December 2027Requirements for high-risk AI systems in areas listed in Annex III apply.
2 August 2028Requirements for high-risk AI systems embedded in certain regulated products apply.

These are legal deadlines under the Act. ISO/IEC 42001 certification has a separate audit process: at Tempo Audits, we assess an organisation’s AI management system against the standard, but certification does not replace its legal obligations under the AI Act.

Where ISO/IEC 42001 Supports EU AI Act Requirements

ISO/IEC 42001 does not reproduce the EU AI Act as a checklist. It gives organisations a management system for identifying and controlling AI risks. Some of its Annex A areas overlap with the Act’s requirements, especially for high-risk AI systems. The connection is useful, but it does not make certification proof of legal compliance.

ISO/IEC 42001 areaOur approachWhat the auditor checks
A.5: Assessing AI impactsRisk management under Article 9 and, where applicable, fundamental rights impact assessments under Article 27Whether the organisation has assessed impacts in the context of the system’s intended use and the people affected
A.6: AI system life cycleRequirements covering technical documentation, logging, accuracy, robustness and monitoringWhether decisions and controls are followed through development, deployment and ongoing monitoring
A.7: Data for AI systemsData governance under Article 10How data is selected, prepared, checked for quality and managed
A.8: Information for interested partiesInformation for deployers under Article 13 and certain transparency duties under Article 50Whether relevant people receive clear, appropriate information
A.9: Use of AI systemsHuman oversight under Article 14 and deployer responsibilities under Article 26How staff use the system, monitor its outputs and intervene when needed

For example, an auditor may trace one AI system from its risk assessment through data decisions, user information and human oversight. That evidence can show how the management system works; it cannot determine by itself whether the organisation has met every legal duty.

Does Your Company Need ISO 42001 Certification?

Yes, if your organisation develops or uses AI and needs to show customers how it manages AI risks. Certification provides independent evidence that your AI management system meets the ISO/IEC 42001 standard.

It is voluntary, but can strengthen customer confidence and support procurement discussions. It does not replace any legal duties your organisation has under the EU AI Act.

Where certification can add value

Customers and business partners may want to know who oversees your AI systems, how you assess risks and what happens when a system’s performance or impact changes. ISO/IEC 42001 gives you a structured way to address these questions and demonstrate that your approach is part of an ongoing management system.

What certification involves

Tempo Audits independently audits AI management systems against ISO/IEC 42001. If the system meets the standard, we can certify it. Our role is to assess the system impartially; your organisation remains responsible for meeting the legal requirements that apply to its AI activities.

Two colleagues working through documents on a laptop

What ISO 42001 Certification Proves and What It Does Not

ISO 42001 certification gives independent evidence that an organisation’s AI management system meets the standard within the scope stated on its certificate. It does not approve every AI system or confirm compliance with the EU AI Act.

What the certificate demonstrates

An audit looks at how the organisation manages AI in practice: who is accountable, how risks and impacts are assessed, how data and system life cycles are controlled, and how people oversee AI use. The evidence needs to show these processes are defined and put into operation, not just described in a policy.

What it does not confirm

The certificate does not establish that every AI system is lawful, safe or compliant with every relevant regulation. The EU AI Act sets obligations according to an organisation’s role and the system’s risk category. High-risk systems may also need separate conformity assessments and technical documentation. The European Commission notes that ISO 42001 helps organisations set up an AI management system, but it does not align fully with the quality management system required by the Act.

How to Get ISO/IEC 42001 Certified: The Process

Preparing for certification can take a few months. Once the Stage 1 audit begins, the audit process typically takes about a month, assuming everything is in order. That usually includes a two to three week gap before Stage 2, followed by three to seven days for us to issue the audit report and certificate.

Stage 1: Review your AI management system

We review your policies and other AIMS documentation, confirm the certification scope and check whether your organisation is ready for Stage 2. If we identify areas that need attention, you have time to address them before the main audit.

Stage 2: Audit how your system works

Tempo Audits assesses how your AI management system operates in practice. This involves interviews, document reviews, process testing and control checks. Once the audit is complete and everything is in order, we issue your audit report and certificate.

EU AI Act and ISO 42001 FAQs

Does using ChatGPT or Copilot at work bring a business under the EU AI Act?

It may. When employees use an AI system professionally under their organisation’s authority, the organisation may be a deployer. Using a third-party tool does not automatically make the business its provider, but duties depend on the system and how it is used. Keep an inventory of workplace AI tools and their purposes.

Are small businesses exempt from the EU AI Act?

There is no blanket exemption for small businesses. The Act still applies according to an organisation’s role and AI use, though some simplified requirements are available to SMEs and were extended to small mid-cap companies. The Commission also says that enforcement should take proportionality into account for SMEs and small mid-cap companies.

Does the EU AI Act require a formal AI training course for every employee?

No specific course, certificate or individual level of AI knowledge is prescribed. Providers and deployers must take appropriate measures to support AI literacy among staff who work with AI, taking their experience and the systems’ context into account. Separate training requirements for human oversight apply to staff overseeing high-risk AI systems.

Can our existing ISO 27001 system support ISO 42001?

Yes. If you already run an ISO 27001 information security management system, the management system structure will be familiar. You may be able to build on shared processes, but ISO 42001 adds AI-specific requirements, including AI risk and impact assessment, lifecycle management and oversight. Your existing certificate does not cover those requirements automatically.

What should we do first to prepare for the EU AI Act?

Start by listing the AI systems your organisation develops, provides or uses, including tools supplied by vendors. Record what each system is used for, who owns it and who may be affected by its outputs. This gives you a practical basis for checking your role, assessing risk and identifying which obligations need attention.

What penalties can a business face for breaching the EU AI Act?

The maximum depends on the infringement. Prohibited AI practices can attract fines of up to €35 million or 7% of worldwide annual turnover, whichever is higher. Other breaches can carry fines of up to €15 million or 3%. These are maximums, not automatic penalties; authorities consider the circumstances of the breach.

Rob Hall

About the author

Rob Hall

CEO, Tempo Audits

Rob Hall is the Founder and CEO of Tempo Audits. A former corporate lawyer and tech entrepreneur, he leads a team providing ISO 27001 and ISO 42001 certification audits and SOC 2 assurance for technology companies.

Book a call

No forms, no faff – just a conversation and a quote. Prefer to skip straight to it? Fill out the application form and we'll get moving.

Alternatively, if you have all the details,
fill out this form here.

ISO 42001