Rollout scope

One team first, or the whole organisation at once

You're a department head with a budget and no mandate to decide for the rest of the organisation, or a compliance lead still working out how wide this needs to go. Both are reasonable places to start. The question is whether your first step makes the next one easier, or harder.

Start with one team

A department head can approve this alone: small budget, no procurement committee, no cross-functional sign-off needed.
You get a working example inside the organisation before asking anyone else to commit.
If the vendor or the approach disappoints, the exposure is one team's budget and time, not the whole organisation's.
Decisions move at the speed of one manager, not a steering group.

Cover the organisation

The article 4 duty follows where AI is used, not which department asked first. If other teams already use AI systems, the same measures-and-context question applies to them too.
One AI register instead of scattered team lists: at /start, sector, AI systems in use, and roles feed a single organisation-wide picture, not five separate spreadsheets.
Depth follows risk per role across the organisation, not one team's own judgment of what matters, so a hiring manager in one department and one in another get the same standard.
When the question gets asked, it's asked at the organisation level: the evidence file org admins export covers the whole organisation, not one team at a time.

See the public log of legal changes

Let the pilot inform the rollout, not become an island

A pilot team is a legitimate way to start, as long as the underlying register isn't built as a team-only artifact. At /start, the intake already asks for sector, AI systems in use, and roles at the level a whole organisation would need, not a scaled-down team version. Run the pilot on top of that structure, and the rollout later means adding teams to an existing register, not rebuilding it. Run it on a separate, team-shaped structure, and the rollout means starting over, this time with a pilot everyone already has opinions about.

When one team really is the right call

Two situations make a team-first decision genuinely correct, not just easier to approve. First, when AI use is truly isolated to one function: a single team using a single tool nobody else in the organisation touches. Second, when a pilot is the internal politics route to a bigger decision: some organisations only move on a broader budget once someone can point to a working example. Neither of these is a compromise. They're the honest answer for a specific situation, not a smaller version of the organisation-wide choice.

What article 4 actually asks for

Since 27 July 2026 (Regulation (EU) 2026/1744), article 4 of the EU AI Act asks providers and deployers to take measures that support AI literacy, taking context into account. It does not guarantee a specific individual level, and it has applied since 2 February 2025, not from a later date. There's no prescribed course, hours, test, or certificate, and no automatic fine tied to article 4 alone. That context requirement is exactly why scope matters: the measures that fit one team's context rarely still describe the organisation's context once AI use spreads beyond it, which is the ordinary case, not the exception.

See what this looks like for your organisation

The 5-minute scan maps your AI use to the roles around it and shows where your evidence is thin.

AI Act evidence routes

Pick the route that matches your AI Act evidence question.

Focused pages per question: from mandatory Article 4 literacy and training records to sector evidence and an audit-ready dossier.

AI literacy in-company training

AI literacy in-company training for organizations, tailored to roles, AI tools, risks, certificates, and Article 4 evidence.

View route

AI literacy program

AI literacy program for organizations: training, course routes, masterclasses, certificates, team reporting, and Article 4 evidence.

View route

AI literacy evidence dossier

The evidence chain from role and use case to scenario practice, scores, certificates, and reporting.

View route

LearnWize versus LMS

An LMS distributes content, LearnWize proves per role that people are ready. Why audit-ready evidence per role beats a generic LMS for EU AI Act Article 4.

View route

Start with one team

A department head can approve this alone: small budget, no procurement committee, no cross-functional sign-off needed.

You get a working example inside the organisation before asking anyone else to commit.

If the vendor or the approach disappoints, the exposure is one team's budget and time, not the whole organisation's.

Decisions move at the speed of one manager, not a steering group.

Cover the organisation

The article 4 duty follows where AI is used, not which department asked first. If other teams already use AI systems, the same measures-and-context question applies to them too.

One AI register instead of scattered team lists: at /start, sector, AI systems in use, and roles feed a single organisation-wide picture, not five separate spreadsheets.

Depth follows risk per role across the organisation, not one team's own judgment of what matters, so a hiring manager in one department and one in another get the same standard.

When the question gets asked, it's asked at the organisation level: the evidence file org admins export covers the whole organisation, not one team at a time.

See the public log of legal changes

Let the pilot inform the rollout, not become an island

When one team really is the right call

What article 4 actually asks for