← Pixel & Paper

&PAPERAI Workshops

AI literacy
you actually experience.

From the first good question to agents and evals, with the two of us at your premises. Pixel brings the tools, Paper brings the dramaturgy. Your team tries things out hands-on instead of sitting through slides, and you get your documentation under Art. 4 of the EU AI Act along the way.

Talk for 30 minutes → The programme
René Weinert (Paper)
Paper
Pixel
Memo Re: Art. 4 EU AI Act

Article 4 of the EU AI Act requires companies to promote the AI literacy of their staff. Since 2 August 2026, supervisory authorities can check this. The obligation does not only apply to AI developers, but to every company in which ChatGPT and similar tools are used in day-to-day work. Even if they were never officially introduced.

Most companies face two problems at once: their staff have been using AI for a long time, just without structure and sometimes with sensitive data. And there is no record that anyone ever explained to them how to do it properly.

The AI licence solves both in a single morning.

Level 1 · AI licence

Five stations.
Hands-on at every one.

The basics for everyone, no prior knowledge needed. No lecture from the front: your team works through station after station, with real tools and tasks from their own daily work. Compact in 2 hours, or as the AI licence in 4 hours with extended exercises and a final check.

  1. How AI ticks

    What happens behind the scenes and where the limits are. Clear, without jargon.

  2. The good question

    The biggest lever of all: turning random answers into usable results.

  3. Wrong, but convincing

    Hallucinations, made-up sources, plausible nonsense. And how to check it in 30 seconds.

  4. Data compass

    Which data must never go into an online tool, and which alternatives exist.

  5. Is it worth it?

    Which tasks AI really takes off your plate, and where it is the more expensive route.

The programme

Five levels.
From the first question to agents.

Modular and freely combinable. Each module takes 2 hours: a short introduction, a live demo, then everyone works on their own. The modules work with the common tools, such as ChatGPT, Microsoft Copilot, Gemini, Claude, Le Chat or an in-house solution. We ask in advance what you use and what is allowed. If no tool has been approved yet, we bring a practice environment.

  1. Basics

    for everyone · no prior knowledge

    How AI works and why it can be convincingly wrong, the good question, which data may go into which tool, spotting made-up answers. The five stations above.

  2. Advanced concepts

    for regular AI users

    Show 5 modules
    Context, model, instruction
    The three levers behind every result. Why the same question gets different answers, how to store background knowledge permanently, and when a different model is the better choice.
    Templates, skills and custom assistants
    Capture good ways of working once instead of rephrasing every time, and share them in the team.
    Connecting AI to tools and systems (MCP)
    How AI searches, calculates, reads files or queries calendars and databases. Which permissions to grant and where the risks are.
    From chat to agent
    What agents can reliably do today, where they fail, and how much autonomy makes sense.
    Which tool, which model, which costs?
    Providers at a glance, cloud or local, subscription or pay per use, data protection.

  3. Evaluating results

    for users and decision-makers

    Show 2 modules
    Checking AI results
    Typical failure patterns, spot checks and cross-checks, verifying sources, and when a human must have the final say.
    Evals: comparing models and solutions systematically
    Your own test cases from real tasks, several models side by side, a scoring rubric. Why a single attempt proves nothing.

  4. Development and automation

    for development, IT and engineering

    Show 4 modules
    AI writes the tool, not the data
    Have scripts generated that run locally on the real data: convert, validate, sort.
    Coding assistants and agents in everyday development
    GitHub Copilot, Claude Code, Cline: plan, implement, get reviews, project knowledge in instruction files.
    Automating workflows
    Chaining AI steps and agents, platforms such as n8n or Dify, and when it pays off.
    AI on your own infrastructure
    Open models locally or in your own data centre: hardware, model choice, costs.

  5. Leadership and adoption

    for management and team leads

    Show 2 modules
    Introducing AI in your company
    Rating use cases by benefit and risk, spotting unmanaged AI use, building blocks of an AI policy, a training plan under Art. 4 of the EU AI Act.
    Adoption workshop (half day)
    Plus work on your own plans: prioritised use cases, next steps, training plan.

Custom workshops

Your tasks.
Your tools.

We work on your team's tasks, with sample material modelled on your real documents. A few use cases that have proven themselves in almost every company:

From meeting to minutes

Transcribe meetings, generate minutes and action items, catch mishearings and invented deadlines.

Asking your own documents

Use manuals, policies and contracts as a knowledge base, with a source reference in every answer.

Research with sources

Market and specialist research with AI, condensing results, checking every claim against its source.

Email and correspondence

Answer enquiries, rewrite texts for different recipients, communicate in several languages.

Analysing documents and spreadsheets

Compare quotes, pull figures from reports, clean up and analyse spreadsheets.

Passing on knowledge

Turn notes and experience into work instructions, onboarding material and FAQs.

We develop further use cases with you, for example for sales, purchasing, HR, quality management or customer service.

Formats and prices

Start small.
Go deeper where it counts.

2 h Basics compact Level 1 with a live demo and a joint exercise. Also any module from levels 2 to 5. up to 15 people from €950
4 h AI licence Level 1 with extended exercises on your own tasks and a final check. Documentation under Art. 4 EU AI Act. up to 12 people from €1,600
1 day Workshop day Three 2-hour modules of your choice, or the AI licence (4 h) plus one module. 6 hours · up to 12 people from €2,400
2 h Custom workshop Your use case, including a preliminary call and tailored practice material. up to 12 people from €1,500

All prices net plus VAT, including preparation, materials and certificates. Online modules from €750. Series discount from five sessions, annual refresher at the module price. Travel within 50 km of Hanover is included, beyond that at cost.

The workshops are held in German. Need it in English? Just ask us.

What you have afterwards

Skills in the team.
Proof on file.

  • Certificate of attendance per person listing content, duration, date and the trainer's qualification
  • Written curriculum as a PDF for your personnel files
  • Attendance list, signed on the day
  • Practical exercise with debrief, not a click-through course
  • Handout: a one-pager for the desk that actually gets used

This gives you the documents for your own record-keeping under Art. 4 of the EU AI Act.

Who runs the session

Two worlds, one workshop.

AI Manager (IHK) · braker.tech

Builds AI systems that run in daily operation, not demos: local language models, agents and automations. He measures new models against real tasks before recommending them.

Fluent in both business and tech: since 2016 he has worked on the development of a web application in a large corporate environment, between business departments, the development team and project management, and runs workshops there himself. Trained IT systems specialist (IHK), Bachelor in business informatics. For him, IT is both job and hobby.

He shows how sensitive data stays in-house: a single powerful computer is enough for local AI today. Whatever can run overnight is finished and checked by the morning.

Covers Live demos · Tools · Data protection

René Weinert (Paper)

PAPER

Founder of anydoors

More than 100,000 participants in team events since 2001. He makes sure listening turns into taking part, and that the workshop feels like a good day, not a mandatory training.

Not a developer but an experienced user: he works with AI every day, from quotes and correspondence to his own apps.

He knows from practice where AI saves time and where you need to look closely. He brings groups together and translates technology into everyday work.

Covers Schedule · Exercises · Practice in everyday work

The next step

One conversation, 30 minutes.

We clarify four things. After that you receive a concrete proposal with a binding quote.

  1. Which tools are you already using?
  2. Who should be trained?
  3. How much prior knowledge is there?
  4. Which tasks would you like to take off your team's plate?
+49 177 6507001 Enquire by email