AI Workshops
Understand AI and use it effectively
From your first prompt to agents and OpenClaw: your team learns how AI works, how to get better results and how to build its own AI workflows.
The right starting point for your team
Start with the basics or focus on more advanced skills. We choose suitable modules, combine explanations with exercises and apply what you learn to tasks from your business.
- Pricing
- From €1,900 net
- Timing
- A single workshop or a progressive series. Modules, duration, dates and group size are agreed after an initial conversation. Preparation and follow-up are part of the agreed scope.
The proposal covers the agreed learning modules and exercises. Licences, API usage and subsequent production rollout are agreed separately.
From the basics to working with agents
Three levels build on each other. You can agree on individual modules, a workshop or a series. The starting point follows your experience and learning goals.
Foundations
Understand AI and write effective prompts
For teams that use AI occasionally or want to start with the fundamentals.
What you can do afterwardsYou can describe a task clearly, provide relevant context and judge whether an answer is useful.
How language models work
How answers are generated, what tokens and context mean, and why plausible claims can be wrong. Learn to distinguish model knowledge, research and reasoning.
View exercise
Generate an answer without supporting material, then repeat with specified sources. Identify which claims are supported, assumed or invented.
Write effective prompts
Define goals, context, constraints, examples and output formats. Use clarifying questions and targeted revisions to improve results and develop reusable prompts.
View exercise
Turn a vague task into a complete brief. Compare two prompt versions against the same criteria and save the better approach as a template.
Manage context deliberately
Select relevant information, break down larger tasks and maintain instructions throughout a conversation. Understand the difference between current context, saved instructions and model training.
View exercise
Turn a cluttered conversation into a concise context pack containing the goal, sources, decisions and open questions. Use it to tackle a new task.
Advanced
Build your own AI workflows
For teams already using AI that want a more systematic and repeatable way of working.
What you can do afterwardsYou can choose suitable tools, connect your own knowledge and develop a workflow with several steps.
Combine models, tools and media
Choose ChatGPT, Claude or Copilot to suit the task. Work with research, files, spreadsheets, images and voice while weighing model choice, source quality and effort.
View exercise
Create a short decision brief from a document and a spreadsheet. Check figures and claims against the original sources and explain your choice of tools.
Use your own knowledge and assistants
Structure projects, custom assistants and knowledge sources. Understand how retrieval-augmented generation (RAG) uses document search and how this differs from training a model.
View exercise
Prepare an assistant with selected documents and clear working instructions. Test missing information and conflicting sources to see whether it attributes evidence correctly.
Turn prompts into a workflow
Connect research, drafting and review in a defined process. Specify inputs, intermediate results, output formats and approvals, then automate suitable steps.
View exercise
Break a recurring task into steps, run through the workflow and identify which handoffs can be automated and where a decision is still needed.
Expert
Work with agents and OpenClaw
For experienced users and technically curious teams that want to develop and understand their own agent workflows.
What you can do afterwardsYou can configure a bounded agent workflow in a test environment, trace its tool use and evaluate it systematically.
Understand agents, MCP and skills
Learn what agents plan themselves, how they use tools and what MCP connections and skills provide. Define tasks, access permissions and stopping conditions.
View exercise
Set the goal, allowed tools and evaluation criteria for a research or coding agent. Turn a recurring instruction into a custom skill and test it.
Use OpenClaw in practice
Understand and configure OpenClaw as an environment for agents: models, workspace, sessions, memory and skills. Try a bounded workflow together in a test environment.
View exercise
Have an OpenClaw agent work with approved documents and create a brief in a test folder. Review its tool calls, context and output together.
Manage quality, cost and control
Compare results using test cases and quality criteria, investigate errors and limit costs. Cover data access, prompt injection, sandboxing, approvals and stopping a run.
View exercise
Test missing sources, misleading instructions in a document and a tool failure. Check whether the agent stops appropriately or asks for a decision.
These levels show the range available. Your proposal specifies the modules and exercises included, the experience needed and the time allocated.
Understand, practise and improve
Each module combines an explanation with an exercise you complete yourself. Your team should be able to apply the method to other tasks afterwards.
One task across three learning levels
From a prompt to your own agent workflow
The same principle is explored as a prompt, workflow or agent exercise, depending on the level. We choose a task relevant to your business.
Example details
- Prompt
- Turn the task, sources and expected result into a clear brief.
- Workflow
- Gather information, create a draft and review it against defined criteria.
- Agent
- Let tools operate within defined access limits and evaluate the workflow with test cases.
Try a method on several tasks
We discuss why an instruction works and what information is missing. You then tackle a new task independently. Comparing results makes mistaken assumptions, unclear requirements and useful improvements visible.
Keep working with your own templates
Exercises produce prompt and context templates, evaluation criteria and, depending on the module, an assistant or workflow. You record what you can change yourself, how to assess results and which questions to explore next.
From a learning project to a production application
The workshop uses exercises and, depending on the module, produces a prototype. Production integrations, access to live systems and ongoing operation are agreed as a separate implementation scope.
AI assistants & automationWhat your team takes away
- An understanding of how AI works and which tasks it can help solve
- Prompt and context templates developed for your daily work
- Depending on the module, an assistant, workflow or tested agent process
- Criteria for reviewing results and a practical plan for continued use
Delivery and handover
Establish experience and learning goals
We discuss team size, current AI use and the skills you want to develop. We then choose modules, exercises and the tools needed.
Understand and practise independently
Short explanations alternate with exercises. You compare results, improve your approach and tackle new tasks independently.
Prepare for everyday use
We save your templates and working examples, address open questions and record how to continue using and developing what you learned.
Before commissioning
Does this fit my business?
Your team wants to understand AI and write effective prompts
You already use AI and want to develop your own assistants and workflows
You want practical experience with agents, MCP, skills and OpenClaw
What do I need to provide?
Team size, experience levels and typical tasks are helpful. We clarify which devices, accounts and sanitised examples are needed for your chosen modules.
Can we start without prior experience?
Yes. The first level begins with how AI works and how to describe tasks clearly. Foundations and prompting need no programming knowledge. Experienced teams can start with workflows or agents.
How much fits into one workshop?
We choose individual modules or a level to suit your goals and experience. Several sessions can be agreed for a progressive learning path. Content, exercises, duration and price are set before you commit.
Does the OpenClaw module include practical setup?
The module uses a prepared test environment or one we set up together. You explore how the model, workspace, sessions, memory, skills and tools interact and try a bounded agent workflow. Prerequisites and required accounts are clarified beforehand.
Which materials does the team keep?
You receive the completed exercises, prompt and context templates, evaluation criteria and next steps. Advanced modules also include the assistants, workflows or skill files developed during the workshop.
Which AI tools and licences do we need?
The selection follows your existing tools and learning goals. We can work with ChatGPT, Claude, Copilot or OpenClaw, for example. Required accounts, devices and access, along with additional licence or usage costs, are clarified before the session.
Where is your team with AI?
Tell me about your experience and what you want to learn or build, from effective prompting to agents and OpenClaw. We use this to plan suitable modules and exercises.