Camp 1 · floor 1: What it really is. Understand in under an hour what an AI assistant really does, what it does not do and why it can be wrong, so you use it without fear or naivety.

Camp 2 · floor 1: The right basics. Use AI simply and well on a real task of the day, with the 3 common-sense reflexes.

Camp 3 · floor 1: Asking well. Get a usable result in three rounds on a real task, by giving a role, a task and a format, examples, and by letting it interview you.

Camp 4 · floor 2: Your workspace. Gather one real work topic in one place (a Project with its instructions, reference files and chats), so the assistant starts from your context instead of from scratch.

Camp 5 · floor 2: The options that change everything. Know which option to switch on for which task (web search, deep thinking, memory, styles, model choice, voice, file analysis) and use it on a real case. Options depend on your plan and your admin: the camp always shows an alternative.

Camp 6 · floor 2: Connect your tools. Plug your email, calendar and drive into the assistant, understand what it sees and what it does not, and ask it targeted questions whose answers you check at the source.

Camp 7 · floor 3: Deliverables, not text. Leave the chat with a finished object you can use as it is (a formatted document, a presentation, a spreadsheet that calculates, a small tool), instead of a block of text to copy.

Camp 8 · floor 3: Chat expert. Finish your chat Second Brain (Project, instructions, connectors, prompt library), judge an output with a grid rather than by instinct, and share a first Project with your team.

Camp 9 · floor 4: Moving to Claude Code. Understand why Claude Code changes the game (it reads and writes files, it runs commands), install it, run a first command and work in a folder, with a no-terminal path for those who will not go there.

Camp 10 · floor 4: Second Brain and Company Brain. Move from the chat Second Brain to the full file version: a written memory, an instructions file read at every session, a GitHub repository, your intentions in writing, one single source of truth. Then open it to the team as a Company Brain, and finally understand connector, MCP, API and plugin.

Camp 11 · floor 4: Advanced automation. Make your recurring work run without redoing it by hand: chain several steps, package a way of working as a Skill, schedule a task that runs on its own, measure the time really saved, and set up the weekly review ritual that keeps the brain clean.

Camp 12 · floor 4: Your first agents. Hand a recurring task to a first dedicated agent, with a job description, guardrails, the human at the right checkpoint and a test of the agent itself. Then see what a team of agents looks like (Atelier Nova, a fictional company, as an example) and sketch your own.

The summit: The certificate waits up here. Each camp you validate raises your level, up to 12 out of 12.

Base camp: Everyone creates their climber here and joins their rope team: their own team. Then one camp after another.

A climber: Everyone goes at their own pace, step by step: read, do, review.

Made it!: Twelve camps later, the certificate. Time to celebrate.

A Sherpa: The pink flag: a colleague who got ahead and helps the others through the camps.

A rope team: Nobody climbs alone: the rope team is your team.

A refuge: A place to catch your breath and share what you found. Smoke from the chimney is a good sign.

The campfire: Those coming down tell their climb. Those going up listen.

The lake: Rumour has it some fish can write prompts. Nobody has checked.

A marmot: She whistles when a climber reaches a camp. Or when she is hungry.

An ibex: The only one here who climbs without an assistant. Respect.

A cow: Not part of the programme. She is here for the atmosphere.

The Tech Nation: The Climb is made by The Tech Nation, in Barcelona.

Alti's snowman: She builds one when you stop moving. The lower it stands, the faster it melts.

A Yeti?: Some say they know every shortcut on the mountain. Nobody has ever caught one.