Academy
Live AI programs, updated to the day they are taught.
AI changes every week. That is why every program is built for one person or one team and updated right up to the day of the class: we teach what works today and take out what has gone stale.
Menu reviewed in
For individuals
Professionals, freelancers, creators and business owners. Three levels and six specialties, on your own work.
See the menu ↓For companies
Teams that want to work with AI every day: assessment, training by area, usage rules and measured results.
See the path ↓Why tailor-made
Always current
We review the menu continuously and before every start. What shipped last week goes into the program; what has gone stale comes out.
On your work
Classes are built on your tasks, your brand or your processes. The examples are yours.
You leave with something running
Every program ends with a deliverable in use: an automation, a published site, a brand system.
For individuals: three levels, six specialties
You start at the level that fits you. On the first call we work out which one it is and which path suits you.
Four paths to find your place
Professional or freelancer
- Foundations
- Real work
- Agents
Business owner
- Real work
- Agents
- AI visibility
Brand or content creator
- Foundations
- Design & content
- AI visibility
No-code builder
- Real work
- Vibe coding
- Agents
These are starting points. The final path is built with you.
Three formats
Any specialty can be taken in any of the three.
One intensive session
Masterclass
We solve one concrete job with a current tool. You leave with one piece running.
4 to 5 weeks
Short program
One full specialty, on your work. You leave with a system in use.
8 to 10 weeks
Full program
The whole path, or one specialty with a guided project of your own. You leave with a project in production.
What every program includes
Live classes
With someone on our team, on your case, with written support between classes.
Every class recorded
So you can watch them again whenever you like.
Materials as files
Guides, prompts and editable templates, organised by class.
What we build stays installed
Skills, automations and projects built in class, in your name.
Current tools sheet
What to use today for each job, with a review date.
Program record
With the detail of what you built.
For companies: from scattered use to a team that works with AI
What makes a team adopt AI is clear rules, managers who use it and access to the tools. The path starts with leadership and ends with measured results.
- 01
Assessment
Which tools exist, how they are used today, and a baseline measurement.
- 02
Leadership session
Ninety minutes that end with priorities and usage rules in writing.
- 03
Kick-off for the whole team
One shared session so everyone starts from the same place.
- 04
Training by area
Four to five weeks on each team's real tasks.
- 05
Internal champions
Two or three people per area who sustain usage and answer questions.
- 06
Processes with agents
We redesign one or two processes with skills, connectors and agents.
- 07
Measurement
Results at 30, 60 and 90 days against the baseline.
You can hire a single step or the whole path. We work on the tools your company already pays for: Claude, ChatGPT, Google Workspace or Microsoft 365.
What we measure
- Active licences per week
- A short practical test, before and after
- Use cases, skills and automations running
- Hours recovered and what they went to
Usage rules, on one page
Approved tools, which data can go into each one, who approves new connectors, and what a person must review before anything is sent, published or paid.
See company training in detail →See audit and consulting for companies →
What we no longer teach
Keeping a program current also means taking things out. This left the menu.
OutAutomations built node by node as the starting point
InAgents, skills and scheduled tasks in the assistant you already use
OutOpenAI's Agent Builder: shuts down on 30 November 2026
InSkills and agents inside ChatGPT
OutSora for video: OpenAI removed its API on 24 September 2026
InKling 4.0, Seedance 2.5 and the current video models
Outllms.txt and special markup to appear in AI answers: Google says they are not needed
InOriginal content, bot access and measurement with first-party data
OutPrompt formulas to copy
InGiving context, asking and reviewing
Build your program
Pick what interests you and book the conversation. It is free and lasts 30 minutes: you can ask every question you need, and we need to understand your case to confirm whether we can help.
Prices are not published because they depend on the format, on how many people take part and on whether the program is taught in Argentina or abroad. After the conversation you get the proposal in writing.
Frequently asked questions
Are the classes recorded?
How much does a program cost?
Does the first conversation cost anything?
Why are the programs tailor-made?
Do I need to know how to code?
Which tools do you work with?
Do you teach in Spanish?
Free guides
Long guides to read right here. No signup: we only ask for your email for the templates.
How to Build a Team of Agents in Codex (working in parallel and reviewing each other)
FreeA main agent splits the work across specialised subagents, each with its own file and its own limits, and returns a single reviewed result. Config files included, ready to copy.
30 minBuildHow to Build a Web App with ChatGPT Sites (from idea to a real URL)
FreeThree apps built end to end: what to write in the prompt, where the data lives, and how to ship without exposing it by accident.
35 minAutomateHow to Create Scheduled Tasks in ChatGPT (that work without you)
FreeReminders, recurring reports and monitors that only ping you when something changes. With ten ready prompts and the protocol to test them.
18 minWorkHow to Create Your Context File (for ChatGPT, Claude, or Whatever LLM You Use)
FreeWrite it once, in twenty minutes. After that every answer comes out with your judgement, your tone and your rules.
25 min