In four weeks, your people ship a working system on their actual work, and the method stays in the company when we leave.
If any of these four is missing, we didn’t finish.
A working tool on a process they already run. Real files, not a sandbox.
A personal AI setup with their standards written in. No more restating rules every morning.
One recurring command a teammate can rerun without the author in the room.
An Agentic Map of what was built, where it lives, who owns it.
Three examples from a single cohort. An IT consulting team that already lived in chat AI. Same method your people would use on their own work.
One client · three artifacts · real processes
Their commercial report sat in a spreadsheet only a few people trusted. In the bootcamp they rebuilt it as a web app, with how the firm counts revenue written into reference docs. They were already using it at work.
From private Excel → shared dashboard
Jobs closed and almost nothing reached marketing. They built a small generator: feed it the client, the space, the scope, the results, and the photos. You get a LinkedIn post, a case study, and a video script in the firm’s voice. A person still checks before it goes out.
From blank page → three assets
Commercial needed call prep, transcript review, and an Ideal Customer Profile that does not reset every Monday. Week one produced the ICP from their own client list, plus a project setup that keeps the context next time.
From start over → reusable agent
Slides for the theory you need, then demo on real work. They watch the loop, then run it on their own work. Last 10–15 minutes are questions.
Reading and exercises, plus setting up their space and hands-on project work. Practical build is the weight we grow over time. Productive work, not busywork.
Managers and directors who already use chat AI and/or Cowork, have hit that ceiling, and will personally build with Claude Code or Codex on a live process they own. People who leave with a working tool, a written setup, a reusable command, and a map — not a briefing.
True beginners (that is the Literacy track). Anyone who only wants a briefing with no hands-on work. Software engineers who already ship with agents. Ops, sales support, coordinators, analysts, and office managers unless they themselves will sit in the build loop.
Learn enough to act, then act. A tight briefing on the idea, then studio time where the agent does real work on their stack.
Before kickoff, a pre-program assessment for every participant. We gauge knowledge, attitude (who is ready vs who will push back and why), map what they actually do day to day, and map where the opportunities sit. This is how we personalize the room before anyone joins a live session.
Live sessions mix slides for theory with demos on real work. They watch the loop, then run it on their own work with us in the room. Length and count follow the format above.
Async between sessions: reading and exercises, plus setup and hands-on project work on their real job. We put growing weight on the practical build. A short follow-up after the program keeps it from fading.
Four weeks. Four things your people can show you. Open a week for what happens in the room and what they take back.
The model is the same for everyone. The plan and the context change the result.
They pick one real process they already run and write what “done” looks like in a sentence a colleague would accept.
We run the same request twice, once with a plan and once without. Then they sort their own work: just do it, plan it, or gate it. Last 10–15 minutes are questions.
They finish a context-rich plan for their own project. Next week they build from it, they do not start from a blank chat.
Short, reversible, no lasting damage if it is wrong. Type it and move.
Multi-step, needs their files and standards, will be reused. Write the plan first.
Customer-facing, money, or a decision you cannot undo. A person signs off before it leaves.
Artifact A context-rich plan for their own project, ready to build.
The chat forgets. The setup remembers. Improving the setup is how they improve the work.
They bring last week’s plan and the files the process actually uses. No sample data.
We build the first working version in front of them, then they run the same loop on their own process. The rules they keep repeating get written into the setup so they are not retyped tomorrow.
They leave a first version that runs on their actual process, plus a setup that already knows their standards.
The setup file is called CLAUDE.md (or AGENTS.md for Codex). It is a short document that sits next to the work and tells the model the team’s standards, files, and “never do this.” That is the whole trick. The chat is disposable. The setup is the asset.
Artifact First working version on their actual process, plus an AI setup that keeps their rules.
By now they have built something real and hit real friction. This week they capture it so the team can rerun it.
They mark the steps they already repeated by hand this week. Those steps are the command, not a new project.
We turn one of those loops into a skill (a reusable command) anyone on the team can run. Then we look at connecting tools they already pay for (email, calendar, the CRM) and when that connection is not worth the friction.
They leave one reusable workflow, plus a written call on what to connect and what to leave alone.
Skills are the reusable procedures they teach the model (a named way to redo a job). MCP is the connector layer: attach tools you already pay for when the round-trip is worth it. If connecting a tool adds more friction than it removes, you do not connect it.
Artifact One workflow they already did by hand, now reusable. Judgment on connecting email, calendar, or the CRM.
The main lesson: working as a team on shared AI work. GitHub is the tool. Review is the habit. The map is what stays.
They get a GitHub account ready and skim how a shared project works: branch, change, review, merge. Plain language, not a developer course.
We walk a shared project end to end on GitHub: the work, the review gate, the handoff. Same idea as a pull request, even if the company does not call it that. Volunteers show a real change; the room gives live feedback. Then each person starts their Agentic Map: what was built, where it lives, how to grow it.
They rerun the loop on their own project and finish the map. Next work starts from that map, not a blank chat.
GitHub here is not a developer ritual. It is how the team shares AI work: someone else can see the change, ask a question, and accept it before it is live. That is how the capability stays when we leave.
Artifact A clear picture of the team review loop on a shared project, plus a personal Agentic Map.
A shareable certificate when they complete the four lessons and the capstone has gone through review.
A report we send you: a diagnostic of automation opportunities surfaced during the program. What to do next, not a survey.
A 1-hour session after a month. What is still running, what needs a nudge, what the next command should be.