Building AI skills
Training only matters if it changes how people work. This module shows how to build AI skills that stick, across every role.
Learning objectives
- Explain the AI literacy obligation
- Map skills needs by role
- Design learning that changes behaviour
- Set up champions and communities of practice
3.1 AI literacy for everyone
Everyone who uses or is affected by AI needs a baseline understanding: what it is, what it's good and bad at, how to use it safely, and how to check its output. Under the EU AI Act, organisations that provide or deploy AI systems must take measures to ensure a sufficient level of AI literacy among their staff (Article 4, applying since 2 February 2025). See AI Literacy for Work and the EU AI Act course.
3.2 Skills by role
| Group | Needs | On this site |
|---|---|---|
| All staff | Literacy, safe use, prompting basics, checking output | AI Literacy, Prompt Engineering, Validating AI Output |
| Managers | Spotting use cases, redesigning work, leading change | This course, AI Business Strategy |
| Leaders & boards | Strategy, risk, governance, oversight | AI Governance, Responsible AI |
| Builders | Integration, evaluation, security, operations | Paths 2 and 3 |
| Risk, legal, HR | Regulation, policy, impact on people | EU AI Act, AI Security |
3.3 Learning that sticks
One-off lectures rarely change behaviour. What works better:
- Hands-on, with real tasks — people practise on their own emails, reports and data (following data rules)
- Short and spaced — 30–60 minute sessions over weeks, not a single day
- Role-specific examples — a prompt library for each team
- Learning in the flow of work — tips, templates and help where people work
- Practice with feedback — share outputs, compare approaches, learn from failures
3.4 Champions networks
AI champions are enthusiastic volunteers in each team who get extra training and time, then help colleagues, collect use cases and feed back problems. A good network has: clear expectations and protected time (for example a few hours a month), regular meet-ups to share what works, a direct line to the central AI team, and visible recognition.
3.5 Communities of practice
Beyond champions, open communities — a chat channel, monthly show-and-tells, a shared prompt library — let good ideas spread sideways across the organisation. Encourage people to share failures as well as wins; they're often more useful.
Design a 90-day AI learning plan for your team: who needs what, the format and timing of sessions, how champions will help, and how you'll know it worked.
See a worked example
Example — 40-person customer service team: Weeks 1–2: everyone completes AI Literacy (self-paced, 3 hours protected). Weeks 3–6: four 45-minute hands-on sessions using real anonymised tickets. 3 champions get extra prompt-engineering training and 4 hours a month. Weeks 7–12: weekly 15-minute tip in the team meeting; shared prompt library. Measures: % trained, weekly active users, handling time, confidence survey.
Key takeaways
- Everyone needs baseline AI literacy — and the EU AI Act requires it of providers and deployers.
- Different roles need different skills.
- Hands-on, spaced, role-specific learning works best.
- Champions need clear expectations, time and recognition.
- Communities spread ideas — including failures.