What it means in practice
An ordinary chat works by you copying text in and the answer back out. Claude Code removes that middle step: you start it in your project folder and the AI sees the files, reads them, changes them, and proposes the changes for your approval.
For a company that means one thing: AI stops advising and starts working.
What we use it for
It is the main tool in both of our technical workshops. Participants build prototypes with it, connect data sources, and have analytics scripts written for them. The level we need from a participant is undemanding:
- activate the right subscription
- install it and sign in
- open the project folder and start Claude Code in it
- write a simple brief
- review and approve the proposed changes
What it takes
A laptop, an active Claude subscription, a GitHub or GitLab account, and basic git: clone a repository, make a branch, commit, push and open a pull request. If git is new to part of the team, we run a short training session before the workshop.
Three days before the workshop, every participant runs a test task that checks Claude Code, git and access all at once. We would rather find out beforehand than on the morning in the room.
A note on tools
We work with tools from Anthropic (Claude), Cursor, OpenAI (ChatGPT) and Google (Gemini). We do not work with Chinese tools, so we do not run workshops on those.