What supply chain evidence do regulators expect under NIS2?
Short answer. Supervisors test supply chain compliance with artefacts, not assertions: a current supplier register, written tiering criteria, dated per-supplier assessments, contracts containing security and notification clauses, a remediation log with owners and due dates, and incident records showing what you knew and when.
Last reviewed: 31 July 2026.
What evidence regulators expect
NIS2 does not contain an evidence list. It does not need to: Article 32 gives supervisory authorities the power to require access to data, documents and information necessary to assess your risk-management measures, and to require evidence of implementation of your cybersecurity policies.
That is the standard. Not "do you have a policy" but "show me the policy operating".
Why evidence is the whole game
Every organisation of any size does some supplier due diligence. The difference between a passing and failing supervisory engagement is almost never the absence of activity — it is the inability to reconstruct it.
Three specific ways good work becomes invisible:
- The assessment happened in an email thread and a personal spreadsheet, and the person left.
- The decision was made in a meeting and never recorded, so the risk acceptance has no owner.
- The evidence exists but takes three weeks to assemble, which is itself the finding.
Article 32(7) also lists the degree of cooperation with the authority as a factor in setting penalties. Slow, partial responses cost twice.
The evidence pack
What a mature organisation can produce within days, not weeks:
How a supervisory conversation goes
The pattern in early-transposing member states is consistent. The authority does not audit the whole estate; it samples.
- "Send us your supply chain security policy and your supplier register."
- "How is the register kept current? Reconciled against what?"
- "We have selected these four suppliers. Send the assessment, the contract and the last review for each."
- "This assessment is dated 2024. What has changed since?"
- "Supplier X had a publicly reported incident in March. What did you do?"
- "Show us where the management body approved these measures."
- "What did you change in the last twelve months as a result of your own findings?"
Steps 3 and 5 are where programmes break. A register that lists suppliers you cannot produce assessments for is worse than a shorter register that is honest about coverage.
Evidence quality tests
Apply these to your own pack before someone else does:
- Dated? Undated artefacts prove nothing about currency.
- Attributable? A named author and approver, not "the security team".
- Traceable? Does the conclusion reference the inputs it was based on, and are those inputs retained?
- Consistent? Does the register agree with the contract, and the contract with the assessment? Contradictions between artefacts are the fastest route to a finding.
- Complete? Does the register match accounts payable and the identity provider, or only the IT budget?
- Retrievable? Can a named person produce it without the original author?
- Versioned? Can you show what the policy said at the time of a decision, not just what it says now?
Common evidence gaps
- Register incompleteness — suppliers procured on cards, through business units, or intra-group.
- Assessment currency — good assessments, two years stale.
- Contract blind spots — the organisation cannot say which contracts contain notification clauses without opening every PDF.
- No remediation trail — findings identified, never tracked, quietly unresolved.
- Unowned acceptances — "we accepted the risk" with no name and no review date.
- No incident timeline — a supplier incident was handled well but nobody logged when notification arrived, so the 24-hour position cannot be reconstructed.
- No linkage to services — the register does not say which of your services each supplier supports, so significance assessments during an incident are guesswork.
Keeping the pack current
Evidence assembled for an audit decays immediately. What holds up is evidence that is a by-product of the process:
- Assessments captured in a system of record, not in documents, so date, author and inputs are inherent.
- Contracts stored with their key terms extracted and queryable, so "which contracts lack a 24-hour notification clause" is a report rather than a project. See adding and managing contracts.
- Screening and monitoring output retained as history, so you can show what you knew on a given date.
- Findings tracked as tasks with owners and due dates, and alerts on overdue items — see managing alerts and notifications.
- A standing report for the management body, produced from the same data rather than reassembled quarterly. Understanding reports covers the reporting views available in Vendorapp.
The test to apply internally: if your competent authority emailed today asking for the four artefacts in step 3 above, how long would it take? If the answer is more than a few days, the gap is operational, not documentary.
Frequently asked questions
- What evidence proves NIS2 supply chain compliance?
- A current supplier register, written tiering criteria, dated per-supplier assessments with their inputs, contracts containing security and notification terms, a remediation log with owners and due dates, and records of incidents and reporting timelines.
- Does a policy document prove NIS2 compliance?
- No. A supply chain security policy is necessary but not sufficient. Supervisors test whether it operates — by picking a supplier and asking for the assessment, the contract and the evidence that the stated process was followed.
- How long should NIS2 supply chain evidence be retained?
- The directive sets no retention period. Retaining assessment history, contract versions and incident records for the life of the relationship plus the limitation period under national law is the common approach — superseded versions matter as much as current ones.