Workshop
AI for Engineering Teams
After the workshop your team works with AI tools the way it worked with an IDE before: as a matter of course, with clear rules and without fear of the repository.
AI in day-to-day development means: an assistant writes, explains and reviews code directly in the flow of work, on request in everyday language. That changes typing less than it changes working: the question is no longer whether the team uses such tools, but under which rules.
Anyone who lived through the introduction of version control knows the pattern: the tool is installed quickly, the shared way of working needs an agreement. This workshop leads through both, from the tool comparison to the written team agreement.
A team agreement on AI tools fits into five lines: what may go into the model and what may not, who reviews generated code, how it is marked, where prompts are stored and what applies in case of doubt. Without these five lines everyone on the team settles it for themselves, and nobody knows how. The workshop ends with this agreement in writing for your team.
Schedule
Agenda
Seven hours including breaks, roughly 40 percent presentation and 60 percent working together on the participants' own code. On request as two half days on consecutive days, so there is time to practise on your own code in between.
Where assistants help and where they do not (45 min)
The three working modes: autocomplete, chat, agent. What each one is suited to and how you notice you are using the wrong one.
The tools compared (60 min)
Claude Code, Cursor, Codex, GitHub Copilot and Gemini CLI side by side on the same task. Differences in context window, tool access, cost and data flow.
Giving context properly (60 min)
Why most poor results are context problems. Project instructions, rule files, selecting the relevant files, working with large codebases.
Live coding: a real task (90 min)
We take a ticket from your backlog and work through it together, from the prompt to the merge request.
Agents and MCP (60 min)
What separates an agent from a chat. Model Context Protocol as the connection to ticket system, repository and internal services. Which tools you give an agent and which you do not.
Rules for the team (45 min)
Mandatory review, what never belongs in the model, handling generated tests, licensing questions, logging. At the end we write a short team agreement together.
Open round (30 min)
Room for the questions that got stuck in day-to-day work, and for topics from your setup that did not come up earlier.
Practicalities
Who it is for, what you bring, what you take away.
Dates and terms are agreed individually.
Contact
Request this workshop.
Write to us which group is to be trained and what should be different afterwards. You get a proposal for scope and schedule.