Most AI training teaches people to write better prompts. That is not the problem. The problem is that your AI knows nothing about your project, and nobody on the team writes that knowledge down in a way it can read. So we spend two days with your team building exactly that, in your own repositories, on your own work.
We have run this both ways, and the difference is not subtle. Send two champions and you get two people who work differently from everyone else. That is DL2 with extra steps, and the gain leaves when they go on holiday.
The files the AI reads before every task have to be written and maintained by the people who know the system. Not by two volunteers, and not by us.
A rule that everybody in the room agreed on holds. A rule that two people brought back from a workshop gets ignored by Friday.
Teams expect the problem to be in coding. It is usually in requirements or testing, which means the people from those phases need to be there.
Two days with everyone beats six weeks of one person trying to convince their colleagues.
No slideware day and workshop day. Both days are working days, and everything we build stays in your repositories when we leave.
We start with where you actually are. Every team scores its own six delivery phases, on the record, no grading and no spin. You see your shape on the six-axis chart within the first hour.
Then we name the bottleneck. Not the general one, yours: the one phase that holds the rest of the team back, with the reason it does.
The afternoon is the framework and the operating model: what DL3 means concretely for your teams, which practices get you there, and what stays out of scope for now.
We build the context system in your repositories. CLAUDE.md as the team brain: your conventions, your architecture rules, the hard constraints the AI must not break. ARCHITECTURE.md and decision records where they are missing.
Then we work through the delivery lifecycle with real tickets from your backlog. Spec-first requirements the AI can execute. Explore, plan, act instead of prompt and pray. Tests generated from acceptance criteria. Release notes from what is actually in the repository.
We close with the plan: what each team owns, what changes in the Definition of Done, and how you will know in four weeks whether it worked.
AI in the IDE is one phase out of six. The teams that get past the plateau put AI to work across all of them, which is what the two days cover.
Spec-first, written so an AI can build from them without a clarification loop.
Design systems and user journeys kept in shared files rather than in someone’s head.
ARCHITECTURE.md and decision records, so the AI proposes changes that fit your system.
The explore-plan-act loop, with the context files doing the heavy lifting.
Tests generated from acceptance criteria, and what to do about false positives.
Documentation and release notes generated from what is in the repository.
Nothing here lives in a slide deck. By the end of day two these exist and are committed.
Two days changes what a team knows. Four weeks changes what a team does. The track is light: we are there for the moments where new habits usually collapse.
The context files meet real work, and get their first corrections.
Specs get written the new way, with us reviewing the first ones.
Review methodology in daily use. We look at the acceptance rate.
What holds, what slipped, and what the next level would need.
Yes, with one condition: cameras on and the full team present for both days. We have run both. The failure mode online is not the format, it is people half-attending from their inbox.
We do not, and we would rather say so. Day 1 gets you the honest read and the framework. Without day 2 nothing gets built, and you have bought insight instead of change.
No. Bring your repositories and your real backlog. We work with the AI tools you already pay for.
That is normal, and it is why every team scores itself separately. The plan comes out per team, not one plan for everybody.
Everyone who ships: engineers, QA, product, leads. If a phase has no one in the room, that phase does not move.
From €8,000, depending on how many teams take part and whether we come to you. The assessment beforehand is free.
Before any workshop we do the 30-minute assessment, so we both know where your teams stand and whether two days is even the right next step. It is free, and it costs you half an hour.