---
title: "NIS2 and CyFun in Belgium: Is Your AI Supplier in Scope?"
url: https://cruxdigits.nl/blog/nis2-cyfun-ai-leverancier-belgie/
language: en
description: "What CyFun and NIS2 in Belgium ask of your AI supplier: the CyFun levels, the 2026 and 2027 deadlines, and the contract clauses that close the gap."
publisher: Crux Digits B.V. (https://cruxdigits.nl)
---

Compliance

# NIS2 and CyFun in Belgium: Is Your AI Supplier in Your Supply Chain?

Your AI supplier rarely falls under Belgium's NIS2 law itself but sits inside your supply chain duty, and the CCB advises such suppliers to meet CyFun Basic. Its model provider is out of reach, so contract for flow-down and notification.

Tom Joseph · AI Consultant 24 September 2026 · 12 min read

NIS2 BelgiëCyberFundamentalsCyFun 2025NIS2 ketenplichtCyFun BasicAI-leverancierCCBVlaanderenmkb+ en enterpriseleveranciersvragenlijstNIS2CyFunCyberFundamentals Belgie

Probably yes, but only if they are your direct supplier. The Belgian NIS2 law puts supply chain security inside your own duty of care, and it reaches the relationship between you and your direct suppliers and service providers. The party holding most of the answers about an AI system, the model provider, usually sits one tier beyond that. Your questionnaire reaches the wrong desk.

This is written for the IT manager, informatiemanager or security officer at a Belgian organisation of roughly 250 to 5,000 employees that has been registered as an essential or important entity, and that has put an AI assistant, a retrieval system or an agent into production alongside an existing ICT partner. You already have a framework choice, an audit calendar and a supplier questionnaire. What you probably do not have is a version of that questionnaire that survives contact with an AI supplier.

One thing first, because Flemish readers are routinely served Dutch material that does not apply to them. The Netherlands has [its own transposition, the Cyberbeveiligingswet](https://cruxdigits.nl/blog/nis2-cyberbeveiligingswet-ai-leverancier/), in force since 15 August 2026. Belgium is a different regime with different dates and a different framework, and it moved first: the Law of 26 April 2024 entered into force on 18 October 2024, making Belgium the first Member State to transpose the directive. If you are reading an article that says your obligations started in August 2026, you are reading about the country next door.

## What is NIS2, and what is CyFun?

NIS2 is the European directive (EU) 2022/2555 on cybersecurity. Belgium transposed it in the Law of 26 April 2024, in force since 18 October 2024. In principle it covers organisations that provide a service listed in Annex I or II of the law, are at least medium-sized (from 50 full-time staff, or an annual turnover and balance sheet total above 10 million euros) and are established in Belgium. They must register on Safeonweb@Work, take eleven minimum security measures, notify significant incidents to the CCB, and have their management body approve those measures. Fines reach 10 million euros or 2% of worldwide annual turnover for essential entities, and 7 million euros or 1.4% for important ones ([CCB](https://atwork.safeonweb.be/nis2)).

CyFun, short for [CyberFundamentals](https://atwork.safeonweb.be/tools-resources/cyberfundamentals-framework), is the framework the CCB uses to make those measures concrete. It has a starting level **Small** for micro-organisations and three assurance levels: **Basic**, the standard measures for every enterprise; **Important**, against targeted attacks by actors with common skills and resources; and **Essential**, against advanced attacks by actors with extensive skills and resources. The route to a CyFun label has four steps: a risk assessment with the CyFun Selection Tool to choose the level, a self-assessment, verification or certification by an authorised conformity assessment body, and the label request on the Safeonweb@Work portal.

For essential entities on the CyFun route the first deadline ran to 18 April 2026, with a Basic or Important verification. An entity whose own risk assessment points to Important has until 18 April 2027 for that verification; one aiming for Essential has until then for the Essential certification. Important entities are in principle supervised only after the fact, but may opt into the same regime and so also gain a presumption of conformity.

## Does the Belgian NIS2 law reach your AI supplier at all?

Not directly, in most cases. The law binds entities that meet a size threshold and provide a service listed in its annexes, using the definitions in [Commission Recommendation 2003/361/EC](https://atwork.safeonweb.be/nis2). An Annex I service at large-enterprise size makes you essential; an Annex I service at medium size, or any Annex II service, makes you important. A twenty-person AI studio building a retrieval system for your document archive is neither.

It reaches them through you instead. Essential and important entities must [secure their supply chain](https://atwork.safeonweb.be/nis2), and the Belgian framing is specific: the duty covers security-related aspects of the relationship between the entity and its direct suppliers or service providers. That word direct is doing real work, and it is the same limit the Dutch law draws. Your AI supplier is in scope of your duty because you contracted them. Whoever they buy their model from is not.

So the enforcement route is commercial, not regulatory. Nobody from the CCB will inspect your AI supplier. An inspector will look at what you asked for, what you were told, and what you did about the gaps.

## Why does the CCB point your suppliers at CyberFundamentals Basic?

Because Belgium built a floor the Dutch regime does not have. The [CyberFundamentals framework](https://cyfun.eu/en/cyberfundamentals-framework-2025), CyFun, is the CCB's tiered control set, and a validated implementation earns a presumption of conformity with the risk-management obligations: until shown otherwise, you are presumed to have met them. The CCB [advises every organisation](https://atwork.safeonweb.be/nis2) that may find itself in a NIS2 entity's supply chain to comply at least with CyFun Basic, and notes that a NIS2 entity could in theory impose a given CyFun level on its direct suppliers.

That advice has teeth now. At the one-year mark the CCB reported [1,500 essential and 2,500 important entities](https://ccb.belgium.be/news/one-year-nis2-belgium-leading-way-and-moving-forward) registered, with 75 percent having already selected a framework and a majority of those choosing CyFun. Four thousand organisations with a framework and an audit date generate a lot of supplier questionnaires, and the ones landing in AI suppliers' inboxes this year are CyFun-shaped.

There was a hard checkpoint on 18 April 2026. Essential entities had to show the CCB inspection service one of [three things](https://ccb.belgium.be/news/nis2-18-april-2026-deadline-what-essential-entities-must-have-place): a CyFun Basic or Important verification obtained or actively in progress, or a signed agreement with an accredited assessment body; or an ISO/IEC 27001 scope, Statement of Applicability and most recent internal audit report, with full certification due by April 2027; or a self-assessment plus a formal request for direct inspection, a route the CCB notes may lead straight to supervisory measures. That date has passed. If you are in scope, the questionnaires are no longer hypothetical.

The inspection service returned to the subject on 11 August 2026, in [a communication to essential entities](https://ccb.belgium.be/news/new-communication-nis2-essential-entities) under article 48 of the NIS2 law. Its theme was the cyber risk arriving with frontier AI systems, and it named supply chain security among the processes that advanced AI changes, alongside governance, risk assessment, monitoring and incident response. The letter went to entities directly rather than being published in full, so treat the summary on the CCB news page as the public record of it. The point here is that the regulator has already put AI and supply chain security in the same sentence.

## Which CyFun questions can an AI supplier actually answer?

More than people expect. Strip the questionnaire down to what a first-tier supplier controls and a competent AI supplier should answer without hesitation:

- How they authenticate into your environment, with which account type, and whether that credential is shared with any other client.
- Which of their staff can read the prompts and documents your users send, under what logging, and for how long.
- Their own basic hygiene: multi-factor authentication, patching cadence, backup and restore, and joiner-mover-leaver for the people on your account.

- Their subcontractor list, and whether any of it changed since the contract was signed.
- The route and the clock for telling you about an incident on their side, which has to fit inside your own reporting obligations.

That last one is where most AI contracts are silent and where the cost of silence is highest. Under the Belgian regime a significant incident requires an [early warning to the CCB within 24 hours](https://ccb.belgium.be/cert/report-incident/nis2-notifications-howto) of becoming aware of it, an incident notification within 72 hours, an interim report on request and a final report within a month. A supplier who tells you on Tuesday about something they saw on Friday has consumed your 24-hour window before you knew it opened. That clause is worth more than the certificate.

## Which questions can only the model provider answer, and will not?

Here is the structural problem, and it is not your supplier being evasive. A questionnaire built for a hosting or software supplier assumes the party you contracted operates the thing you are asking about. With an AI system, four of the most load-bearing questions are about infrastructure your supplier rents:

- What happens to a prompt after it leaves your supplier's application: how long it is retained, whether it enters abuse monitoring, and whether a human can read it.
- Which region a specific inference call actually ran in, as opposed to which region the account is configured for.
- How much notice, if any, is given before a model version changes underneath a stable API.
- Whether a model update counts as a change your supplier is contractually obliged to tell you about.

A model provider publishes terms for all customers and negotiates individually with very few. Your supplier cannot make those terms say something else, and a small AI supplier has no more influence with a frontier lab than you do. Pressing harder on the first tier produces a longer answer, not a better one.

So the honest verdict is that a supply chain duty stopping at the direct supplier stops one party short of the answers. That is not a drafting flaw to argue around. It is the shape of the obligation, and the useful response is to stop asking the first tier to warrant facts it does not control.

## What do you put in the contract instead of a certificate?

Three things, and none of them requires your supplier to hold a CyFun label they were never going to buy. First, a flow-down obligation: the supplier names the model providers and subprocessors behind the service, keeps that list current, and tells you before it changes. You are not making them responsible for the upstream party's security. You are making them responsible for the transparency you cannot get yourself.

Second, notification symmetry. The supplier tells you about an incident, a material terms change, or a model version change within a window that leaves your own 24-hour early warning intact. Write the hours down. A clause saying "without undue delay" does not survive contact with a Friday afternoon.

Third, and this is the one most often skipped, a named change category for model behaviour. A model that starts answering differently without a deployment is not an availability event, and it will not appear on anyone's monitoring. We have argued elsewhere that [nobody gets paged when an AI system is up and wrong](https://cruxdigits.nl/blog/ai-incident-ownership-ict-partner/); the contractual version of that argument is that if model behaviour is not a change, nobody has to tell you it changed.

What you are building here is a record that you asked, were told, and acted. Under ex-ante supervision for essential entities, that record is the thing an inspector can see. A certificate your supplier does not hold is not.

## Does CyFun 2023 or CyFun 2025 apply when you send the questionnaire?

Both, for now, and it is worth getting right before you send anything. The CCB [released CyFun 2025](https://ccb.belgium.be/news/cyfunr-2025-here) in October 2025, aligned more closely with the NIST Cybersecurity Framework 2.0 and with NIS2. Its headline changes are directly relevant here: more focus on supply chain security, more focus on operational technology, and governance measures added from the Important assurance level upward.

The CCB confirms on its [CyFun page](https://atwork.safeonweb.be/tools-resources/cyberfundamentals-framework) that CyFun 2023 and CyFun 2025 will coexist until 18 April 2027, and that entities choose which version to apply until then. Accredited assessment bodies, in their own transition guidance, keep 2023-based statements valid until 18 April 2028 at the latest; treat that second date as their reading and confirm it with your own assessor. The planning consequence holds either way: if your own conformity work is heading for CyFun 2025, ask your suppliers the 2025 supply chain questions now rather than re-asking them next year.

Note also that you may [justify a CyFun assurance level below your NIS2 classification](https://atwork.safeonweb.be/nis2) on the basis of a risk analysis, and that the inspection service can sanction you for conforming to a lower level than your risk actually warrants. The same logic should govern what you demand of an AI supplier. A Basic-level expectation on a supplier whose system touches your customer records is a decision you will have to defend, not a saving.

## Who pays for CyFun preparation in Flanders?

When a customer imposes CyFun Basic on a Flemish supplier, the question is who pays for the preparation. Cybersecurity advice through the kmo-portefeuille is no longer possible: VLAIO announced on [15 October 2026](https://www.vlaio.be/nl/nieuws/budget-kmo-portefeuille-is-uitgeput) that the budget is used up, and the scheme ends on 1 January 2027. VLAIO itself points to the [cybersecurity verbetertrajecten](https://www.vlaio.be/nl/subsidies-financiering/cybersecurity-verbetertrajecten): an approved provider coaches the company, and VLAIO subsidises 50% of trajectories between €7,100 and €39,000 excl. VAT for SMEs and maatwerkbedrijven, and 35% for companies that are not SMEs but fall under NIS2. At the time of writing VLAIO lists nineteen approved providers.

That is not a route for the AI supplier to pay for its own CyFun label, but it is a way for a Flemish SME in your chain to reach Basic without carrying the whole cost. The wider picture of what remains after the kmo-portefeuille is in [AI subsidies for Flemish SMEs](https://cruxdigits.nl/blog/ai-subsidies-flemish-smes-vlaio-2026/).

## Where this leaves an IT organisation that already has partners

You do not need a new ICT partner to do any of this, and we are not offering to be one. Crux Digits [builds and integrates AI systems](https://cruxdigits.nl/applied-ai-ml-mid-sized-business/) alongside the ICT partner or internal IT team you already have; we do not run service desks, infrastructure or estate management, and we are not a conformity assessment body, so we cannot verify anyone against CyFun. What an AI supplier owes you here is answerable evidence about its own half of the system and honesty about the half it rents.

The practical order of work is unglamorous. Take your existing supplier questionnaire, split it into what the first tier controls and what it rents, get firm answers on the first half, get the flow-down and notification clauses on the second, and keep the correspondence. If the AI system also [falls under the AI Act](https://cruxdigits.nl/eu-ai-act-risk-checker/), the obligations run in parallel rather than replacing each other, and the reporting clocks are not the same length.

If you want the supplier questions [settled before go-live](https://cruxdigits.nl/ai-implementation-netherlands/) rather than after the first audit letter, that is a conversation worth having while the architecture is still movable. For how that works in a Belgian project, from audit to handover, see [AI consultant in Belgium](https://cruxdigits.nl/ai-consultant-belgium/).

Talking to an AI consultant in Flanders?

A consultant tells you where AI pays off; Crux Digits also builds it. A fixed price per step, one named expert, and the Belgian rules and funding built in.

[AI consultant in Belgium →](https://cruxdigits.nl/ai-consultant-belgium/)

## Frequently asked questions

### What is the difference between CyFun Basic, Important and Essential?

They are the three CyFun assurance levels. Basic holds the standard measures for every enterprise, Important adds protection against targeted attacks by actors with common resources, and Essential against advanced attacks by well-resourced actors. Basic and Important are verified, Essential is certified, each by a CCB-authorised conformity assessment body.

### Is an ISO 27001 certificate from our AI supplier enough?

It helps, but it answers a different question than the one you are being asked. ISO/IEC 27001 is one of the two recognised routes for your own conformity in Belgium, provided the scope and Statement of Applicability cover the measures at a level equivalent to your applicable CyFun level. A supplier's certificate is evidence about that supplier's management system, not a presumption of conformity for you, and it says nothing about what a model provider upstream does with a prompt. Read the scope statement before you accept it: a certificate scoped to a head office says little about the platform running your workload.

### Do we have to report an AI system's wrong answer to the CCB?

Only if it meets the definition of a significant incident, which most wrong answers do not. The Belgian law defines an incident as an event compromising the availability, authenticity, integrity or confidentiality of data or of the services offered, and significance turns on severe operational disruption, financial loss, or considerable material or non-material damage to others. A model producing a poor answer within normal operation is a quality problem, not a notifiable incident. A model producing wrong answers because its retrieval source was tampered with is a different matter, and that is exactly why the distinction belongs in your triage rules before it is needed.

### We are not in scope for NIS2. Can a customer still impose CyFun on us?

Yes, and that is the most common way a Flemish company first meets the framework. The CCB itself notes that an organisation outside the law's scope can be pulled in through a contractual requirement because its customer carries a supply chain duty, and advises such organisations to meet CyFun Basic as a minimum. Separately, the CCB can identify an organisation as essential or important regardless of size in specific circumstances, for instance where it is the sole provider of a service. Being small is not by itself an exemption from either route.

### Does the Dutch Cyberbeveiligingswet apply to our Flemish operation?

Not to an entity established in Belgium. Both laws transpose the same directive, but each Member State's law binds the entities established there, and an entity with establishments in several Member States is subject to each transposition separately, with the national authorities cooperating on inspections and incident notifications. If your group has a Dutch establishment too, that part faces the Dutch regime and its own dates. The practical consequence is that a group-wide supplier questionnaire written to one law will under-serve the other.

### Who inside our organisation signs off on this?

The management body, and that is not a formality in the Belgian regime. Management bodies of NIS2 entities must approve the cybersecurity risk-management measures and oversee their implementation, and they are liable if the entity breaches those obligations. Members are also obliged to follow training sufficient to identify risks and assess risk-management practices. In practice that means an AI supplier decision with a known, documented gap needs to be a decision someone at that level has seen, not a note in an IT architect's file.

Our AI services [AI consultancy](https://cruxdigits.nl/ai-consulting-netherlands/) [AI automation](https://cruxdigits.nl/ai-automation/) [AI agents](https://cruxdigits.nl/ai-agent-development-netherlands/) [AI implementation](https://cruxdigits.nl/ai-implementation-netherlands/) [Pricing](https://cruxdigits.nl/pricing/)

Related reading

[NIS2 AI Supplier: Who the Law Actually ReachesUnder the Dutch Cyberbeveiligingswet, three parties around one AI system have three different legal positions. Why...](https://cruxdigits.nl/blog/nis2-cyberbeveiligingswet-ai-leverancier/)[AI Supply Chain & Demand Forecasting Guide for ManufacturersAI supply chain optimisation for manufacturing enables sharper demand forecasting, leaner inventory and earlier sup...](https://cruxdigits.nl/blog/ai-supply-chain-demand-forecasting-guide/)[AI Adoption in Belgium 2021 to 2025: From 7th to 4th in the EUWe pulled five years of Eurostat survey data on AI use in Belgian companies and compared it with the Netherlands an...](https://cruxdigits.nl/blog/ai-adoption-belgium-2021-2025-data/)[AI for Housing Corporations: Sector Standards FirstYour housing corporation's IT estate already runs on CORA, five Aedes data standards and a thirty-supplier covenant...](https://cruxdigits.nl/blog/ai-woningcorporatie-datastandaarden/)

## Want any of this applied to your business?

We turn these concepts into working tools: grounded, safe and measurable. Start with a free consultation.

[Book a free consultation →](https://cruxdigits.nl/#contact)
