An AI Rollout Without Frustration: Why the Organisation Counts Twice as Much as the Tool

Buy licences, send an email to everyone, done. That is how many AI rollouts go. Three months later, a fifth of the staff uses the tool regularly; the rest gave up after a disappointing first try.

What the data says

For the 2026 Work Trend Index, Microsoft surveyed 20,000 AI usersU in ten countries, including Germany. The central finding: organisational factors explain 67 % of AI impactU, individual factors only 32 %U. Whether AI delivers in a team depends more on leadership, rules and room to experiment than on individual talent. Yet only 13 %U of respondents are rewarded for experimenting with AI.

Germany is moving. According to Bitkom, 25 % of companies using AIU offer no training, down from 43 % a year earlier. And the Danish study by Humlum and Vestergaard shows that where employers train staff and provide an official tool, workers report larger time savings.

Training is also a duty

Since 2 February 2025, Art. 4 of the AI Act requires companies that use AI to take measures on their staff's AI literacy. The Omnibus Regulation rewrote the duty in July 2026: what is required are measures to promote AI literacy, not a guaranteed level per person. Documented training therefore makes sense twice over.

The workflow: a rollout programme

Champions are the most important building block. They know their team's tasks and show what works in real day-to-day work. Give them time for it, or the role becomes a volunteer job on top of their actual job.

Limits and risks

• Training without a use case: Generic prompting courses fade fast. Train on the tasks the team actually has.
• Pressure instead of freedom: People afraid of making mistakes do not experiment. Explicitly allow trying things out.
• Measuring success: Weekly active use is the early indicator; measured time savings are the real value.

How to start

1. Write the rules on one page and publish them.
2. Name one champion per team with a fixed time budget.
3. Plan three role-based trainings with real everyday tasks.
4. Record participation and content as evidence for Art. 4 of the AI Act.

Sources

1. UGeekWire: Microsoft's new research finds an AI paradox holding companies back, 2026
2. UForbes / Moor Insights: Microsoft Work Trend Index 2026 Shows AI Productivity Is Not Enough, 19 May 2026
3. UBitkom: For the first time, a majority of companies use AI, 14 Sep 2026
4. SHumlum, Vestergaard: Large Language Models, Small Labor Market Effects, NBER WP 33777, 2025
5. RLaw and Technology: AI literacy, the Digital Omnibus rewrites Article 4 of the AI Act