Back to Blog

AI Training Obligation Under the EU AI Act: What Your Team Needs to Do Now

Henri Jung, Co-founder at Superkind
Henri Jung

Co-founder at Superkind

Superkind logo on an abstract grey waves texture with soft warm light

AI training obligation under the EU AI Act means this: since 2 February 2025, every company that uses AI must take measures to support AI literacy among its staff1. In plain language, your team needs to know what the KI-Agent does, where it makes mistakes and when it asks for approval - before something goes wrong. This guide shows what Article 4 requires, who needs which training, how to document it, and how Superkind’s ready-made AI team turns the obligation into a running process across your tools.

Key takeaways

Get a ready-made AI team - Superkind comes with an AI team for AI literacy (Literacy Planner, Policy Guide, Compliance Watch) that runs automatically. Start for free with unlimited users, or build your own.

Keep control, keep data in Europe - Superkind is hosted in Europe, GDPR compliant, never trains on your data, and you decide which steps need approval.

Article 4 applies now - since 2 February 2025, companies using AI must support staff AI literacy. No certificate required, but documentation is expected.

Role-based, not one-size-fits-all - sales, finance and HR each need different training content. The Bundesnetzagentur recommends four steps: determine needs, develop measures, refresh, document4.

Superkind already covers much of it - transparent logging, defined routines with approval steps, and a clear audit trail mean your team sees what the AI does at every point.

Your AI team for AI literacy

OnboardingPeople
Literacy PlannerAI employee
New employee in Personio
New in Personio this week: 2 people. AI basics training and policy sent. 1 confirmation received, training log in Excel updated.
Training done28 of 30
OpenL. Meier, starts Monday
Next policy review1 April
Runs automatically for every new hire
Example: Literacy Planner at work in Microsoft Teams

What is AI literacy?

AI literacy is the ability to understand how an AI system works, what data it uses, where it can go wrong and when a person needs to step in. Article 4 of the EU AI Act (Regulation (EU) 2024/1689) makes it a legal obligation: companies that deploy or use AI must take measures to support this understanding among everyone who works with the systems1. Superkind is an example of an AI agent that makes this concrete: you see what it does in Outlook, Teams or Slack, you set the approval rules, and the log shows every step - so the team learns by working with it, not by reading a slideshow.

How it works with Superkind: your AI team for AI literacy

You write to Superkind by email, in Teams, in Slack or in the web app, like a colleague, and Superkind does the task in your programs: via 1,700+ integrations, including Personio, Notion and Excel. It works with your existing tools instead of replacing them, hosted in Europe and GDPR compliant.

With Superkind you don't start from zero: it comes with ready-made AI employees for AI literacy and compliance. They run automatically in the background, and you decide where they ask you first. Use them right away, adapt them or build your own.

Literacy Planner

AI employee for AI literacy documentation

The EU AI Act expects you to support AI literacy, but nobody remembers who completed the training. When a new person is added in Personio, the Literacy Planner sends the training materials and the policy, collects the confirmation and keeps the training log in Excel up to date. Every month you get the list of open items.

Works inMicrosoft TeamsPersonioExcel+ 1,700 more
OnboardingPeople
Literacy PlannerAI employee
New employee in Personio
New in Personio this week: 2 people. AI basics training and policy sent. 1 confirmation received, training log in Excel updated.
Training done28 of 30
OpenL. Meier, starts Monday
Next policy review1 April
Runs automatically for every new hire
Example: Literacy Planner at work in Microsoft Teams

Policy Guide

AI employee for AI policy questions

Nobody reads the full AI policy before pasting data into a tool. The Policy Guide answers questions in #ai-questions straight from your policy in Notion: which tool is approved, which data never goes in, who signs off. If the policy has no answer, it says so and tags the person responsible.

Works inSlack+ 1,700 more
#ai-questions
Lena Weber
Can I upload the quarterly revenue report to a free AI summariser?
Policy GuideAI employee
New question in #ai-questions
Not to a free tool. Section 2 of the AI policy: financial data only goes into approved tools with a DPA. Upload it to Superkind instead, it summarises in Europe.
Policy section2 - Approved tools
Approved routeSuperkind, hosted in Europe
Runs automatically for every question in #ai-questions
Example: Policy Guide at work in Slack

Compliance Watch

AI employee for AI Act compliance checks

Art. 4 of the AI Act says you need to support AI literacy, but nobody checks whether the documentation is complete. Every quarter the Compliance Watch pulls the training log, compares it with the team list in Personio, and sends you a report with gaps, expiring trainings and new roles that need a different module.

Works inOutlookPersonioExcel+ 1,700 more
Q4 AI literacy report
Compliance WatchAI employee
Routine - first Monday of the quarter
Quarterly AI literacy check done. 28 of 30 people trained. 2 gaps: L. Meier (new, starts Monday) and T. Fischer (role changed to sales, needs the customer-data module).
People trained28 of 30
Open gaps2
Next check1 April
Runs automatically every quarter
Example: Compliance Watch at work in Outlook

Every employee is ready to use right away. Or build your own on the platform.

Safe by default

AI literacy starts with the tools you approve. With Superkind, the approved tool already runs in Europe, keeps your data out of model training, logs every action and follows your approval rules - so your team sees exactly what the AI does.

What you need

Before you start

  • One person who owns the training plan (team lead, HR or the person responsible for data protection)
  • A list of the AI tools your team uses, including private accounts
  • Your AI policy - if you do not have one yet, see our step-by-step guide
  • A place for the training log everyone can access: Excel, Notion or SharePoint
  • A free Superkind account at app.superkind.ai, connected to your systems via 1,700+ integrations such as Personio, Slack and Asana (up to 150 hours of work free, then from EUR 50 per month for unlimited users, as of November 2026, see pricing)

Step by step with Superkind

Step 1: Map your AI tools and roles

List every AI tool in use, who uses it and for what. Include free tools and private accounts - the Bundesnetzagentur expects you to cover all AI systems your staff interacts with3.

SlackDaniel to Superkind

Ask in #general which AI tools people use for work, for what task and with which account. Collect the answers in a table by Friday and save it in Notion under AI Literacy.

Step 2: Define training content by role

Not everyone needs the same training. Someone in sales who uses an AI agent for follow-ups needs to understand customer-data rules; someone in finance needs to know why certain numbers require a human check. The Bundesnetzagentur recommends matching the depth to the task and the risk4.

RoleWhat they need to knowExample
EveryoneWhat AI is, data classes, approved toolsWhich data never goes into a free chatbot
SalesCustomer-data rules, approval for quotesQuotes over EUR 10,000 need sign-off
FinanceWhen to verify AI output, audit trailCheck totals before posting to the ledger
HRPersonal data, bias awarenessNever feed CVs into unapproved tools
ManagementOversight, documentation duty, review cycleSign off the quarterly compliance report

Step 3: Run the first training and document it

Keep it short and practical: 30-60 minutes per role, real examples from your own tools, and a test question at the end. The important part is the documentation. Article 4 does not prescribe a certificate, but the Bundesnetzagentur recommends a log with name, date, topic and confirmation3, 4.

TeamsDaniel to Superkind

Send every team member the AI basics training link and the AI policy. Collect confirmations. Update the training log in Excel with name, date and topic.

Step 4: Automate onboarding for new hires

Every new person needs the training before they start using AI tools. The Literacy Planner picks up new hires from Personio, sends the materials and tracks the confirmation - without you remembering to do it.

Step 5: Set a review cycle

AI tools change, roles change, the law changes. Set a quarterly review: the Compliance Watch compares the team list with the training log and flags gaps before anyone asks.

Set it up as a routine

OutlookDaniel to Superkind

On the first Monday of every quarter: compare the team list in Personio with the training log in Excel. List gaps, expiring trainings and new roles, and send me a report.

  • Every new hire - training materials and AI policy go out automatically from Personio, confirmation lands in the log.
  • Every quarter - a compliance report: gaps, new tools, roles that changed.
  • Every six months - review and update the training content. Superkind keeps the new materials in your Company Brain and uses them from that point on.

As your compliance needs grow - separate modules per department, assessments, a dashboard for management - you build your own app on the Superkind platform, on top of what Superkind already knows. It belongs to you.

Data protection and EU hosting

  • Approval where you want it - you decide which steps wait for your OK before Superkind sends, books or changes something. Everything else runs on its own.
  • Sovereign and independent - hosted and processed in Europe under European law, models from OpenAI, Anthropic, Mistral and Google processed in Europe, no lock-in to one model vendor.
  • Transparent logging - every action is logged. Your team sees what the AI did, which data it used and where the result went. That is AI literacy in practice.
  • DPA ready - sign the data processing agreement through the Trust Center.

Other ways

WayEffortError riskBest for
SuperkindLow: set up once, runs every hireLow: logged, tracked, gaps flaggedTeams that want compliance on autopilot
Spreadsheet and remindersMedium: someone updates itMedium: gaps when people forgetSmall teams with a dedicated owner
External training providerHigh: cost and schedulingLow on paper, open in practiceRegulated industries, formal certificates
Do nothingZeroHigh: no proof if something goes wrongNever a lasting answer

Common mistakes

  • One training for everyone - sales and finance need different examples. Role-based content is the standard the Bundesnetzagentur recommends4.
  • Training without documentation - if you cannot show who was trained and when, Article 4 is hard to demonstrate to an authority.
  • Ignoring private tools - a free chatbot on a personal account still falls under the AI Act when used for work.
  • Training once, never again - tools change, roles change. Set a review cycle or your documentation goes stale.
  • No approved alternative - banning AI without offering an approved tool pushes use into the shadows. Name the approved path in the training.

6 tips for AI literacy in your team

1. Start with the AI policy

The AI policy is the foundation. Training without clear rules teaches nothing people can follow.

2. Use your own examples

“Never paste the revenue report into a free chatbot” sticks better than “avoid sharing confidential data with unapproved AI systems”.

3. Let the tool teach

Superkind’s transparent logging, approval steps and audit trail mean the team learns by working with it. Every action shows what the AI did and why.

4. Check the DPA first

For every tool with personal or financial data, check the DPA, data location and training opt-out before the first upload. See our sovereign AI explainer for what to look for.

5. Document as you go

A log with name, date and topic is enough. The Literacy Planner does this for you with every new hire and every quarterly check.

6. Review after every new tool

A new AI tool means a new module. Our European AI tools guide helps you pick tools that already meet the bar.

Conclusion

Article 4 of the EU AI Act does not ask for a 200-page compliance manual. It asks you to support AI literacy - and to prove it. Map your tools, train by role, document every step, and review it quarterly. Superkind takes three of those steps off your plate: it sends the training, collects the confirmation and flags gaps across Personio, Slack and Outlook via its 1,700+ integrations. Your first step: ask your team which AI tools they use today.

Frequently Asked Questions

Article 4 requires providers and deployers of AI systems to take measures that support AI literacy among their staff. Since 2 February 2025, every company that uses AI in the EU must ensure people working with AI understand what it does, where it makes mistakes, and how to use it responsibly. The Digital Omnibus (Regulation (EU) 2026/1744) softened the wording from "ensure" to "support", but the duty remains.

Article 4 does not carry its own fine. Penalties are set by national authorities under Article 99. In practice, missing documentation weakens your position if something goes wrong: a data-protection authority or court can treat it as an aggravating factor. The Bundesnetzagentur recommends documenting all measures so you can show compliance at any time.

The AI Act does not prescribe a certificate or a fixed format. The Bundesnetzagentur suggests four steps: determine needs by role, develop fitting measures, refresh regularly and document everything. Measures can be training sessions, internal guidelines, coaching or self-paced courses - what matters is that the level matches the task and the risk.

Superkind turns the training plan into daily practice: the Literacy Planner sends training materials to every new hire from Personio, collects confirmations and keeps the training log in Excel. The Compliance Watch checks gaps every quarter. It works in your programs through integrations with Personio, Notion, Outlook and many more, and natively via computer use where no integration exists.

Superkind is hosted and processed in Europe, GDPR compliant and aligned with the EU AI Act (Regulation (EU) 2024/1689). It does not train models on your data, and you sign the data processing agreement (DPA) through the Trust Center at trust.superkind.ai. AI models from OpenAI, Anthropic, Mistral and Google are processed in Europe.

Yes. Someone in sales who uses AI to draft follow-ups needs different knowledge than someone in finance who hands invoices to an AI agent. The Bundesnetzagentur recommends role-based measures. The key is proportionality: the training matches the task and the risk of the AI system the person works with.

Related Articles

Henri Jung, Co-founder at Superkind
Henri Jung

Co-founder of Superkind, the AI employee teams write to in Outlook, Teams and Slack. Henri helps companies, from small teams to the Mittelstand, hand their routine work to AI - without new tools, consultants or an in-house AI team.

Send Superkind your first training task.

Start for free in Outlook, Teams or Slack. Or let us show you in 30 minutes how it runs in your systems.