AI literacy & governance
Article 4 requires trained staff. Enforcement starts 2 August 2026.
Most companies using AI already carry this duty and have no record to show for it. The workshop produces the record.
The obligation
Every company using AI is in scope
Article 4 of the EU AI Act has applied since 2 February 2025. It requires providers and deployers to ensure a sufficient level of AI literacy among staff and anyone operating AI on their behalf, including contractors. National market surveillance authorities begin enforcing it on 2 August 2026.
This is not only a high-risk-systems duty. If your staff use ChatGPT, Copilot, or any AI feature inside the software you already run, you are a deployer and Article 4 applies. The 2026 simplification package postponed several high-risk deadlines to December 2027, but it did not move the literacy obligation, which makes it one of the first duties authorities can act on.
Sources: EU AI Act Article 4 (official) · European Commission AI literacy Q&A · Council, 7 May 2026
- Who it covers. Employees, contractors, and service providers who use or operate AI on your behalf.
- What "sufficient" means. Training matched to each role and to the risk of the systems that role touches, not a single all-staff video.
- What counts as proof. The Commission's guidance is clear that reading the tool's instructions is not enough; you keep internal records of who was trained, on what, and when.
- What enforcement looks like. Authorities can require corrective action and, under the Act's penalty regime, fines up to €15 million or 3% of global annual turnover for breaching its obligations (Article 99).
The offer
A role-based workshop that leaves a record
Delivered remote or on-site, in English, French, or Hungarian.
Board
Obligations, risk acceptance, and the decisions a director actually signs off on.
Legal & compliance
Classification, the deployer duties, and where AI rules meet GDPR.
Operators
Tool-specific oversight for the people who run AI systems day to day.
General staff
The rules for everyday AI use, and how to recognise and report a failure.
You walk away with the evidence pack
The deliverable is the documentation an authority asks for: attendance records, a role-mapped curriculum showing who was trained on what, and assessment results. It is the literacy record your own AI policy should treat as an audit artifact.
- Attendance records per session and per role.
- Role-mapped curriculum with versions, so you can show what was taught.
- Assessment results that demonstrate the training landed.
- Optional add-on: a governance starter, the five published artifacts adapted to your company.
Proof
We publish our working standard
Five complete worked examples of the documents every EU deployer needs. Read them before you decide.
AI system inventory
The survey found 6 systems. The SSO logs found 9.
AI policy
Two prohibited practices still being sold to companies.
Risk assessment
Why the rights assessment hits the pricing model, not the hiring tool.
Model card
Honest documentation includes the amber findings.
Incident response plan
The artifact authorities ask about second, after training records.
All five →
The Resources hub, with the free PDFs and the editable bundle.
Pricing
Fixed scope, one-page proposal
Indicative, in the same style as the tabletop tiers. The final quote follows a short scoping call.
A single role-based literacy session with its training record.
- One audience (board, legal, operators, or staff)
- Curriculum matched to your AI footprint
- Attendance and assessment record
Role-based training across the organisation, with the full evidence pack.
- Workshops for every role tier
- Role-mapped curriculum and versions
- Complete evidence pack for your records
- Annual refresh option
The five governance artifacts, adapted to your company.
- Inventory, policy, risk assessment
- Model card and incident response plan
- Built on the published worked examples
How it fits
Literacy is the floor. Exercises are the stress test.
Training and documentation get you compliant and give your people the basics. A tabletop exercise then tests whether the decisions actually hold when an AI system fails. Many clients do both: the literacy programme first, the exercise once the governance is in place.
- Literacy programme. Trained staff and the evidence pack. The Article 4 floor.
- Tabletop exercise. Your team rehearses a real AI-failure scenario and finds where the response breaks.
- Together. Documented capability, then a tested response chain.
Get the literacy record in place before 2 August
A short call to scope your AI footprint, the roles that need training, and the evidence you'll need to show.