If you’re working towards ISO 42001 certification, you need a clear, practical guide to building an artificial intelligence management system (AIMS) that both functions efficiently and stands up to audit scrutiny.
That’s what our team has put together on this page. At ISO QSL, our ISO consultants help organisations like yours develop ISO-compliant management systems. This guide walks you through ISO 42001 implementation step by step, from initial planning through to getting ready for your certification audit.
Are you prepared for ISO 42001?
We’re here to help you review how your organisation uses AI, identify gaps against the standard and prepare for ISO 42001 certification.
Book a discovery call
About the author
Jodie Turner – Marketing Team Leader
During my time as ISO QSL, I’ve developed extensive knowledge of digital marketing alongside a strong understanding of the ISO standards that help organisations improve.
What’s the goal of ISO 42001 implementation?
ISO 42001 is a full management system built around governance, risk and operational control of your artificial intelligence (AI) systems.
Your ISO 42001-compliant AIMS should define how your organisation uses AI, identify risks and impacts, put controls in place across the AI lifecycle, and monitors, audits and continually improves your system.
In other words, your aim is to build a working system of policies, processes and controls that improves the efficiency, reliability and governance of your organisation’s AI.
Step 1: Secure leadership commitment
Like all ISO standards, ISO 42001 starts at the top. Without leadership involvement, your implementation will almost certainly stall or fail at audit. You need to discuss and define clear ownership of your AIMS, align it with your broader business objectives, and define everyone’s roles and responsibilities.
Step 2: Define scope and AI system boundaries
Next, define the scope of your AIMS. This will be different for every organisation. For example, it may include business units, products or services covered by AI, AI systems or your organisation’s role in relation to AI (e.g., provider, deployer, etc). Overly broad scopes create unnecessary complexity, while overly narrow scopes lead to unnecessary audit risks.
Step 3: Carry out a gap analysis
Before building a new system, assess where you are now. A gap analysis compares your current practices against the ISO 42001 requirements. If you already have some form of AIMS, you’ll likely already have some elements in place, even if they aren’t structured or documented correctly. Focus on your governance structure, risk management processes, AI lifecycle controls and documentation.
Step 4: Identify AI systems and lifecycle processes
As part of your AIMS development, you need to understand how your AI systems operate across their full lifecycle. This involves identifying all in-scope AI systems and recording them in a register, mapping how they’re designed, developed, deployed and monitored, and defining the control measures at each stage. This stage can help you identify any gaps in visibility, especially in larger corporations.
Step 5: Policies and processes
Next, your system should include your AI policy and objectives, your governance framework, and defined processes for managing your AI systems. These core policies provide a solid foundation onto which you can build the rest of your system. If you already hold ISO certifications like ISO 9001 or ISO 27001, or you’re working towards them, you may be able to integrate some elements from these rather than starting from scratch.
Step 6: Perform risk and impact assessments
Risk and impact assessments are central to ISO 42001, or indeed any ISO standard. You must assess the specific risks associated with each AI system, the potential impacts on individuals, organisations, society, and any other stakeholders, and the likelihood and severity of those impacts. Importantly, this should go further than any traditional IT risk assessments, such as those you may have carried out for ISO 27001.
Step 7: Implement controls and governance processes
Now, based on your AI risk assessment, you implement controls. These should typically include data governance and quality controls, model validation and monitoring, human oversight, traceability and documentation. You’ll find a structured set of controls in Annex A of ISO 42001 which you can use to address identified risks. Not every control will apply to every organisation, which is why you should apply these in the context of your organisation, rather than ‘copying and pasting’ them in.
Step 8: Assign responsibilities
Your AIMS will only work if people understand what they need to do. Ensure it contains responsibilities specifically assigned to teams or individuals, as well as leadership (as mentioned previously). Your aim here is to develop a clear accountability system across your entire organisation, reflecting real-life responsibilities.
Step 9: Train staff
Once your people know their roles, they need training. Take an individual approach. Everyone should understand the importance of the ISO 42001 AIMS and their role in it, but tailor the training. Deliver general awareness training for your wider teams, role-specific training for those managing AI systems, and leadership training for those in charge.
Step 10: Monitor, audit, and review the system
Before you achieve ISO 42001 certification, you need to prove your system works. This involves monitoring your AI systems’ performance and taking corrective actions, conducting internal audits, and carrying out management reviews. When it’s time for your audit, auditors will expect to see this evidence.
Step 11: Prepare for ISO 42001 certification
By this stage, you should have a good understanding of your AIMS. The certification process typically begins with an initial assessment of your management system, followed by a certification audit. Your chosen certification body will explain their specific certification process and assessment requirements.
If you’re working with a non-accredited provider like ISO QSL, you don’t necessarily need to have everything in place before starting the certification process. Many organisations engage ISO consultants during implementation to help develop their AIMS, address gaps in their management system, and prepare for certification. The right support can help you build confidence and avoid common pitfalls during the implementation process.
Do you need a consultant to implement ISO 42001?
You can absolutely implement ISO 42001 internally. Many organisations try. However, because ISO 42001 is a relatively new standard, many of the standard’s requirements are easy to misunderstand or misinterpret. This often leads to unclear sections, or missing requirements entirely. In turn, your audit may flag some issues, and you may not achieve certification at that audit.
Overcomplexity is a significant issue we’ve seen many organisations run into. ISO 42001 doesn’t have to be complex. In fact, the simpler you can make it, the easier it is for your staff to understand. As a result, they’re more likely to implement it properly in their day-to-day operations.
Working with an experienced ISO consultant can significantly speed up your implementation, reduce the amount of rework you may have to do, and improve your audit outcomes. At ISO QSL, we support organisations at every stage of their ISO 42001 journey. Whether you’re starting from scratch, integrating ISO 42001 with existing standards, or working towards certification.
If you’re starting your ISO 42001 journey, or you need to make changes quickly to get ready for an upcoming audit, our team can support you at every step. Get in touch with us today for a no-obligation conversation about what we can do for you.