
You map overlapping requirements by working from one control set, not from three laws. NIS2, DORA and the EU AI Act each add obligations, but the majority of their requirements land on the same controls: risk management, access control, incident handling, supplier oversight and documentation. Build the mapping once — control by control — and each new law becomes a delta on what you already have, not a new program. Euraika GRC keeps these mappings built in: 458 cross-framework mappings across 30 frameworks, so a control you implement once satisfies every framework it maps to.
Why does every new law feel like starting over?
Because each regulation arrives with its own vocabulary, its own authority and its own checklist — and organisations respond by building a parallel compliance track for each. But the underlying asks converge. Directive (EU) 2022/2555 (NIS2) requires cybersecurity risk-management measures and incident reporting. Regulation (EU) 2022/2554 (DORA), applying to EU financial entities since 17 January 2025, requires ICT risk management, incident reporting and third-party oversight. Regulation (EU) 2024/1689 (AI Act) requires risk management, data governance, logging and human oversight for high-risk AI systems; since its 2026 amendment (Regulation (EU) 2026/1744), those duties apply from 2 December 2027 for the systems listed in Annex III and from 2 August 2028 for those covered by Annex I.
Three different regulators, largely the same controls. An incident-reporting process built for NIS2 is most of an incident-reporting process for DORA; a risk register built for DORA's ICT framework is most of a risk register for the AI Act's system-risk assessment.
Which requirements actually overlap?
From the texts, the shared core is: risk management (NIS2 Art. 21, DORA Art. 6–9, AI Act Art. 9), incident handling and reporting (NIS2 Art. 23, DORA Art. 17–19, AI Act Art. 73), access control and security measures (NIS2 Art. 21, DORA Art. 9, AI Act Art. 15), supplier and third-party oversight (NIS2 Art. 21(2)(d), DORA Arts. 28–30, AI Act supply-chain duties), and documentation and auditability (all three, everywhere). What genuinely differs: the AI Act's conformity assessment and CE-marking track for high-risk systems, DORA's resilience testing, and NIS2's management-body accountability regime. Those are the deltas — the parts you cannot reuse.
How do you build the mapping in practice?
Row by row, control by control:
- Start from one baseline framework. ISO 27001's Annex A or, in Belgium, the CyberFundamentals Framework (CyFun) gives you a neutral control vocabulary that every regime can map onto.
- For each control, list which article of which law it satisfies. This is the mapping. Access control → NIS2 Art. 21(2)(i), DORA Art. 9(4)(c), AI Act Art. 15. Do it once and every audit afterwards reads from it.
- Tag evidence by control, not by law. A pen-test report serves all three regimes when it is filed against the control it tests. Filed against "the DORA audit", it serves one.
- Handle the deltas explicitly. AI Act conformity assessment, DORA threat-led penetration testing, NIS2 management training — give each genuinely new obligation its own row so it cannot silently hide behind an overlapping one.
What does an AI Act obligation look like inside an existing program?
The AI Act's high-risk obligations apply from 2 December 2027 for Annex III systems, and most of them slot into existing structures rather than replacing them. An AI inventory is an asset inventory with extra fields (purpose, risk tier, deployer and provider roles). AI-system risk assessment extends the risk register you already keep. Logging of AI-system decisions is audit logging with a defined scope. Human oversight is a role assignment inside the access-control matrix. The genuinely new part is the classification itself — deciding which of your systems are high-risk under the Act's Annex III — and that is a legal question more than a technical one.
When is it worth automating the mapping?
Doing the mapping by hand is feasible for one organisation and two or three laws — it is a few weeks of expert work, and then a maintenance burden that grows with every new framework. It stops being feasible when the mapping must stay current: laws amend, control frameworks update, and a hand-maintained spreadsheet drifts. That is the point of Euraika GRC: the 458 cross-framework mappings are maintained in the platform, evidence is tagged to controls automatically, and each new law lands as a delta rather than a rebuild.
Frequently asked questions
Do I need the same controls for NIS2, DORA and the AI Act?
Not the same controls, but a shared core. Risk management, access control, incident reporting, supplier oversight and documentation overlap across all three. Only each law's specific regime — DORA's resilience testing, the AI Act's conformity assessment, NIS2's accountability rules — needs dedicated work.
Which framework should be my baseline?
A recognised control framework you can be audited against: ISO 27001, or in Belgium the CCB's CyFun. The baseline matters less than the discipline of mapping every law onto one vocabulary instead of running parallel checklists.
How often must the mapping be reviewed?
At each legal change: implementing acts, guidance from the national authority, the ESAs for DORA, or the AI Office for the AI Act. A mapping is a living document; annual review is a reasonable cadence between changes.


