For youth work and social business
Everyone on the team uses AI. Nobody has rules for it.
We bring your team to a shared level, develop rules with you that people actually follow, and build the first use case together. Afterwards you know what you use, why, and with which data.
Why now
The usage is there. What is missing are skills, a framework and guidance.
- 91%
- of 12 to 19 year olds in Germany use AI applications (JIM study 2025)
- 73%
- of staff and volunteers in the non-profit sector use generative AI (ZiviZ 2025)
- 9%
- of non-profit organisations have binding rules for AI (ZiviZ 2025)
Young people are not waiting for youth work to catch up. Professionals are already experimenting, mostly on their own. What turns individual trial and error into good practice is a team that understands AI, agrees on rules and keeps the relationship with young people where it belongs: with people.
Sources: mpfs, JIM-Studie 2025, p. 61 f.; ZiviZ im Stifterverband, “Individuell genutzt, aber strategisch vernachlässigt”, December 2025, pp. 4 and 8.
What we offer
Three building blocks. Book them separately or combine them.
Every organisation starts somewhere else. Some teams need the basics first, others already have use cases and need rules. We put the blocks together with you.
AI training: basics and deep dives
From a first understanding to real depth in individual models. The basics give the whole team a shared footing. The deep dives go into Claude and ChatGPT in detail: projects, skills, memory, connectors, agents. Everyone develops a use case from their own everyday work.
Who it is forProfessionals and teams, ideally with leadership in the room. Basics for everyone, deep dives for those who are further along.
Show units (8)
Basics
Digital change and youth work: figures, facts, questions
60 min
Current figures on how young people and organisations use AI, and what they mean for your field.
How language models tick: basics, limits, hallucinations
90 min
Live demos show how a model produces text, why it sounds confident when wrong, and how to check results.
Tool marketplace: five AI applications compared
120 min
Stations with ChatGPT, Claude, Gemini, Mistral and Perplexity: the same task on each, strengths and limits noted.
Data protection and law in brief: GDPR, EU AI Act
60 min
What you may enter where, what the AI Act requires of your organisation, and how to work with practice data.
Deep dives
Deep dive Claude: projects, skills, memory, connectors, Claude Code
half day
Set up a project with its own knowledge, reusable skills and memory, and connect it to your tools.
Deep dive ChatGPT: projects, custom GPTs, data analysis, voice
half day
Build a custom GPT for a recurring task, analyse a spreadsheet and try voice mode for everyday work.
Advanced prompting: roles, meta prompts, review loops
90 min
Let the model improve your prompt, give it a role and a checklist, and build a review step into every workflow.
Your own use case: from prototype to next step
half day
Everyone builds a first version of their own use case, gets feedback in the gallery and leaves with a next step.
- You leave with
- A team with shared basics, real depth in one or two models and first use cases of its own.
- Format
- One to four days, on site or online, by arrangement.
AI guidelines workshop
Develop shared rules for using AI in your organisation: what for, with which tools, with which data, who checks results. Data protection and the transparency duties of the EU AI Act included.
Who it is forLeadership, board, data protection lead and interested professionals.
Show units (4)
Stocktake: who already uses what, and what for
90 min
An honest picture of current AI use in the team, including the grey areas nobody has talked about yet.
Guardrails: fields of use, data, transparency, human review
half day
The team agrees where AI may be used, which data stays out, how AI use is disclosed and who checks results.
Tool approvals and data protection check
90 min
Which tools are approved for which tasks, based on their data protection terms and your own requirements.
Decision paper for board or leadership
60 min plus review round
The results become a short guideline draft with a decision proposal, reviewed once with the team before it goes up.
- You leave with
- A draft guideline your board or leadership can decide on.
- Format
- Two workshops plus a review round, by arrangement.
Practice lab: build systems, work together
Turn one idea into a working system: an assistant for enquiries, a knowledge base with connectors, an agent structure for funding applications. And we design how your team works with it: shared projects and knowledge, clear roles, who checks what, how results are handed over. After four to eight weeks we review together what holds up in everyday practice.
Who it is forSmall implementation teams of two to six people, with a clear use case.
Show units (6)
Sharpen the case: benefit, users, data
60 min
We clarify who the system is for, what it should do and which data it may and may not use.
System design: knowledge base, connectors, roles and review steps
half day
On paper first: which knowledge the system needs, which tools it connects to, who checks what before anything goes out.
Build and test with practice data: assistant, agent or workflow
half day
The team builds the first working version and tests it against realistic cases, with practice data only.
Working as a team with AI: shared projects, shared knowledge base, who maintains what
90 min
How the team shares projects and knowledge in the tool, and who keeps them up to date.
Team workflows: handovers, approvals, documentation, quality checks
90 min
Where AI results are handed over, who approves them, how they are documented and how quality is checked.
Follow-up after four to eight weeks: what works, what sticks
60 min, online
A look back at what the team is actually using, what got stuck and what the next step is.
- You leave with
- A working system, a plan for how the team works with it and a follow-up conversation.
- Format
- Lab day plus follow-up session, by arrangement.
How we work
Enablement, not dependency.
Your team builds the solutions.
We facilitate the process and bring the AI expertise. The use cases, the rules and the decisions come from your team. That is what makes them stick.
Data protection first.
We work with practice data, never with real data about young people. We choose tools deliberately and look at where data goes before anyone types.
Relationships stay human.
AI can prepare, sort and draft. Listening, advising and being present for young people stays with people. AI takes work off the plate so there is more time for exactly that.
Funding in mind.
Digital transformation is a priority of Erasmus+ and the European Solidarity Corps. We show where your prototypes connect to these programmes.
From our practice
Four days of AI training for JUGEND für Europa
- 4
- days in Stuttgart, 29 Sep to 2 Oct 2026
- 18
- professionals from youth work, sport and civic engagement
- 9.8/10
- would recommend the training to colleagues
Working with Bianca was a real pleasure. We experienced the joint preparation and delivery of the training ‘AI in Erasmus+ Youth, Erasmus+ Sport and ESC – Actively Shaping Digital Transformation’ as very enriching in terms of content, professional and built on trust. We would particularly like to highlight the flexibility and composure with which the trainer duo responded to short-notice needs and challenging situations. The very positive feedback from our participants fully confirms this impression. We look back on a thoroughly successful training.
Patrick Stähler and Katharina Erbes
Programme Officers, JUGEND für Europa
German National Agency for Erasmus+ Youth, European Solidarity Corps and Erasmus+ Sport
Where does your organisation stand?
In a short intro call we look at where your team is today and which building block fits now. No strings attached.
Mingle Collaboration is the AI brand of Linie 22 GmbH. Bianca Rohrbach has worked in facilitation and organisational development for more than 20 years and leads the AI trainings.