AI-LIT-LAB · Path 1: Understand and use safely
AI Literacy Hands-on Lab
Half a day after which every participant has built their own assistant instead of just hearing about one.
- Duration
- ½ day
- Group
- up to 12 people
- Format
- In-house · Remote · In the lab
- Language
- German or English
- Price
- €3,950 flat, plus VAT
Who this track is for
Teams and departments that need to move from listening to doing: case handling, sales, administration, service. Ideal as the second step after the AI Literacy Praxislabor (AI-LIT-LAB), or as a direct entry point if the basic vocabulary is already in place.
Who it is not for
Not for teams that already work with assistants daily and want to improve their actual workflows. They belong in Working Safely and Productively with AI (AI-WORK).
Starting situation
After a briefing or first private experiments, many people know what a language model is, but hardly anyone has operated one under guidance. Some do not dare, others paste customer data into a private ChatGPT account. Between knowing and doing safely, a protected practice space is missing.
This exists after the track
- Every participant has written, improved and evaluated their own prompts on a real task from their own daily work
- The data rules are practised on concrete cases: every participant has decided and justified at least ten classifications themselves
- A first own assistant with role, tone and limits, set up on our lab environment or your tenant
- A passed transfer check per person as a documented building block of your Art. 4 competence evidence
Prerequisites
Basic terms such as prompt and hallucination should have been heard before, for example in the AI Literacy Praxislabor (AI-LIT-LAB). A laptop with a browser is enough, prior AI experience is not required.
Preparation before the track
Every participant brings a task profile: one recurring task from their own daily work, described in three sentences, without personal data. This exact task is what the lab works on.
What's included
- Half-day hands-on lab with Dino Bordonaro, on site or remote
- Prepared practice accounts on our lab environment, or setup on your tenant by prior arrangement
- Exercise set with data rule case cards and prompt starter templates
- Step-by-step guide for rebuilding the assistant in your own organization
- Attendance certificates and audit-proof training documentation including transfer check results
Agenda
From knowing to doing
Every participant writes the first prompt for the task from their profile. The results are compared: Why does the same question produce twenty different answers? This starts the first improvement round.
Data rules under fire
Case card exercise: customer complaint, salary list, press release, quote calculation. Each participant decides per case whether and into which AI the data may go, and justifies the decision. Contested cases are argued out in plenary.
Prompts that work repeatably
The morning's attempts are distilled into a pattern: role, task, context, format. Each participant reworks their own prompt along the pattern and measures against the output what actually improved.
Your first own assistant
Each participant sets up an assistant with their own role description, tone and clear limits, on our lab environment or your tenant. By the end, the assistant answers the profile task better than the first prompt of the morning.
Transfer check and Monday plan
Each participant demonstrates their assistant on a new case and decides three data rule cases in the check. Closing question: What do I rebuild on Monday, and what do I need from whom?
Exercises and lab share
Around three quarters of the time, participants work at the keyboard themselves: their own prompts, their own case decisions, their own assistant. The trainer demonstrates, corrects and deliberately provokes failures.
Platforms
The default is our prepared lab environment without any customer data. On request we practise directly on your Microsoft tenant in whichever version is available to you, in which case we clarify access and approvals with your IT in advance.
Transfer evidence
Transfer check at the end: each participant demonstrates their own assistant on a new case and decides three data rule cases. Results and attendance are documented in an audit-proof manner.
Artifacts you take home
- The self-built assistant including its role description, ready to rebuild in your own organization
- Personal prompt starter set along the pattern role, task, context, format
- Data rule case card set for internal refresher exercises
- Training documentation with transfer check results for the compliance archive
Optional extensions
- Working Safely and Productively with AI (AI-WORK) to improve real workflows
- Prompt and Context Engineering (AI-CONTEXT) for heavy users and power users
- Additional lab sessions in the same week for further teams
Boundaries
The hands-on lab is practice with documented evidence. It replaces neither tool selection, nor tenant setup as a project, nor policy work (AI-GOV).
Frequently asked questions
Do we need our own licences or tenant?
No. By default everyone practises on our prepared lab environment with practice accounts. If you want to practise on your own tenant, we clarify licences and access with your IT beforehand.
Is half a day enough to build your own assistant?
Yes, because nothing is explained that is not practised immediately, and because every participant works on their own profile task. The assistant is simple by the end, but real and rebuildable. Deepening the work on full workflows follows in AI-WORK.
What data is used in the exercises?
On the lab environment, exclusively practice data and the participants' own task profiles without personal data. The data rules are not just a topic, they apply in every exercise, and violations become teaching moments.