<aside> 🧭
A one-day, on-site workshop for one cross-functional team of four to eight people. We set up a practice company that resembles yours, with a working AI-native environment inside it. Your team runs a real project in it. You get a one-page diagnosis of what has to change in your own organisation. If you prefer, the same day runs on your existing project, repository or databases.
</aside>
Teams that have tried AI assistants and now want to make them part of daily work. The workshop is built for mixed groups: engineering, operations, product, quality, administration. A technical background is not required.
The one requirement is that everyone on the team has used an AI assistant before. First contact with a language model should happen before the workshop.
| Duration | One day |
| Location | On-site, at your company |
| Group | Four to eight people from one company, cross-functional |
| Format | In-house only, hands-on, with live support during the project work |
| Tools | Model-agnostic. The patterns work across Claude, GPT and Gemini |
| Delivered by | Stellar Work forward-deployed engineers |
This part is our work. We prepare a practice company that looks like yours: the same types of documents, the same kind of legacy systems, the same kind of projects. A pharma R&D team gets a pharma lab. An operations team gets a plant.
We also build the AI-native environment your team will work in: knowledge a model can read, retrieval over the documents, reusable instruction sets and small automated workflows. That way the day itself goes to using it.
Your team works on realistic material from the first hour, and no confidential company data is needed in the room.
The practice company is the default because it needs no access to your systems. If your team would get more out of working on the real thing, we run the same day on your existing project, repository or databases.
The structure stays the same. Before the workshop we build the environment over your own material. On the day your team runs a piece of work from your actual backlog in it.
| Practice company | Your own project | |
|---|---|---|
| Material | A company we set up to resemble yours | Your existing project, repository or databases |
| Access needed | None | Access to the material in scope, arranged before the workshop |
| Company data in the room | None | Yes, within the scope we agree beforehand |
| What stays afterwards | Patterns and know-how to transfer | An environment that already sits on your real work |
| Preparation | A short conversation about your setup | The same conversation, plus scope, access and data rules agreed with your IT or security contact |
In the table below, read "practice company" as "your project" if you choose this option.
| Block | What happens |
|---|---|
| Orientation | The practice company: what it does, what its legacy looks like, and how the prepared environment is set up |
| Project brief | A project brief mapped to the real jobs of the people in the room |
| Project work | The team runs the project inside the environment we prepared |
| Live support | We step in when something breaks and explain why: why the wiki is structured that way, why retrieval returned this result and not another |
| Retrospective | What worked, what broke, and what would have been different at home |
Result of the day: a one-page operating-model diagnosis and a friction report, handed to you and the team.