AI TRAINING
Microsoft Copilot training: 55 employees in 9 sessions
Industrial SMB
Category
AI training
Industry
Manufacturing
Duration
9 sessions over 2 months
Trained
55 employees
Key result
8.5/10 satisfaction, 67% regular users
The context
An industrial SMB wanted to train all 55 of its employees in generative AI and Microsoft Copilot. The goal was ambitious: build a shared AI culture, help every participant make their first productivity gains, and lay the foundations for responsible, secure use.
I designed the training program and coordinated 2 trainers, Nassim Zga and Sonja Toussaint, to work with 55 employees split into 9 sessions of 8 to 12 people over 2 months.
What I did
Morning: discovery and introduction (3.5 hours)
Going around the room to list repetitive tasks (broken down with the STEP method). An introduction to generative AI and Copilot: LLMs, assistants, agents, workflows, MCP. A live Copilot demo on the participants' own computers. A security workshop in small groups: risks (confidentiality, hallucinations, deepfakes) and setting 3 golden rules. An introduction to prompting with the CRAFT method (Context, Role, Action, Format, Tone). A “Briefing the assistant” role play in pairs. A demo of high-impact use cases.
Afternoon: prompt-a-thon (3.5 hours)
Group ideation: picking 8 to 12 quick wins from the time-consuming tasks identified in the morning. Work in small groups: each group takes 2 use cases and produces the best series of prompts to solve them. The trainer acts as a “Copilot coach,” moving between groups. The focus is on testing and iterating.
Sharing and contributing (1 hour)
Each small group presents its solution, and the whole room celebrates each group's discoveries. Launch of the central Teams channel and the shared prompt library (OneNote/SharePoint). Immediate action: the groups add their prompts to the shared library live, becoming its first contributors or building on the work of earlier groups.
Coordination over 2 months
9 sessions scheduled and coordinated with 2 trainers. Each session adds to the shared prompt library, and the last groups benefit from the work of the earlier ones. Adoption is tracked after the training.
The results
8.5/10
overall satisfaction (53 respondents)
98%
would recommend the training
67%
use AI at least once a week
Skills progress
| Skill | Before | After | Progress |
|---|---|---|---|
| Understanding generative AI | 3.9/10 | 7.1/10 | +80% |
| Understanding AI agents | 1.2/10 | 5.2/10 | +316% |
| Understanding automation | 2.3/10 | 3.8/10 | +70% |
| Readiness to integrate AI | 4.2/10 | 6.6/10 | +57% |
What participants took away
Saving time and broadening the scope of research on technical topics
Understanding how to write a more effective and relevant prompt
A better understanding of the different types of tools
Organizing my work, time savings and quality
Want similar results?
Training 55 employees takes a proven program that works in the room and holds up on compliance. The training course on responsible AI with Copilot applies this format to your teams.
30 minutes, no commitment.
