01
It takes your input
Files, notes, a brief, a spreadsheet — whatever your task actually starts with.
In person · Lisbon · groups of 11
Leave the room with an AI agent that already does work from start to finish.
In 4 hours you build an AI agent for a real task from your own job and walk out already using it.
No coding required
What you build
An agent that finishes the job
By the end of the session your agent runs one complete process on its own, start to finish.
02
It applies your rules
The way you do the work, saved as skills the agent follows the same way every time.
03
It reaches your services
Connected over MCP and API, so it can fetch what it needs and act beyond its own folder.
04
It saves the result
A deck, a spreadsheet, a document — finished, in the place you keep your work.
Why now
A year of prompting, and nothing accumulated
You have used AI chats for months. The reason none of it turned into a tool is structural, not a matter of writing better prompts.
A chat
- Starts from zero every session
- Forgets your rules the moment the tab closes
- Cannot reach your files or your services
- Leaves the assembly work to you
An agent
- Keeps your rules as skills and reuses them
- Lives in a project that persists between sessions
- Connects to your services over MCP and API
- Hands you the finished artifact
Programme
The 4 hours and 5 blocks
30 min
How an agent is built
How an agent differs from a chat, and how to give it a task you can check.
55 min
MCP connectors
Your agent acts beyond its own folder.
55 min
Skills
Your agent follows rules you saved once.
40 min
The context architecture and project folders
A project you can pick up again tomorrow.
50 min
Connecting over API
One full process, input to finished artifact.
3 h 50 min of work. With breaks, the 4 hours.
Included in the price
4 h
In-person workshop
Lisbon, a group of 11. You build step by step alongside Egor.
+1
Group call a week later
Online. We go through what you have built since and answer your questions.
+5
Video lessons
5 more lessons that go deeper into questions the 4-hour session does not cover.
What you keep
What you take home
4 things sitting on your own machine, and 1 you carry in your head.
01
A configured project
Set up in Codex, ready to open again tomorrow.
02
Working skills
Several of your own — instructions that became repeatable processes.
03
Connected services
MCP connectors wired to the tools you actually work in.
04
Live API connections
Services connected by key and tested on your own task.
05
The method
How to hand your agent a new process, step by step, on your own. The only intangible one — and the one that keeps paying.
Bring one task. Leave knowing how to build it, with a concrete plan.
See the datesBefore you come
What to bring
A laptop
Charged, with an adapter. Any operating system.
A paid ChatGPT subscription
About €20 a month. Claude works too — just tell us before the session so we can prepare. Set it up before you arrive: we start building in the first 30 minutes.
About €5 of API credit
Spent during the session on your own connections.
One real task
Something from your actual job that you repeat and dislike. Your agent gets built around it.
Roughly €25 in total, on top of your ticket.
Fit
Who this is for
Come if
- You already use AI chats and know you want to go further with AI
- You have a repeating task you would happily never do by hand again
- You are comfortable installing things and following steps live
- You want to keep building agents yourself afterwards
Skip it if
- You are looking for a programming course — there is no code here
- You want to watch rather than build; this is 4 hours of hands-on work
- You need an agent built for you rather than by you
Who runs it
A practitioner, not a theorist
Egor Maraev
Founder of Robot & Sons · Created an AI-agents course with 100+ graduates · 13 years running businesses
A founder for 13 years: first a video production company, then a computer graphics studio, now Robot & Sons. Running his own businesses, he had to automate a great deal of the work himself, so he teaches AI implementation as a practitioner, from what already runs in his companies.
LinkedIn ↗Reviews
What Egor's students say
This course was the best money I have spent in a long time. I started using automations right after the first session.
Eva Kozyreva
AI creator, freelance producer
My production now has its own skills for SMM, articles, editorial checks, visuals, proposals and production documents. Codex handles briefs, estimates, Google Docs and the routine that used to eat hours.
Vladimir Miloserdov
Owner, MILOSERDOV production
You did not just organise what I knew about agents — you did it for my own field. It is far more interesting to learn when the examples come from your work and you can apply them the same day.
Irina Finn
CEO, PYATNITSA Studio
I definitely recommend this course: it is compact and extremely useful. You start applying what you learn straight away, and the calls with Egor answer any question on the topic.
Yuri Zaytsev
Film director
A wonderful course, one of the most useful I have ever taken.
Alina Shalimova
Producer, Legko Studio
After the first session I had solid ground under my feet, and after the second, real results at work.
Maxim Kalinichenko
Director, head of production, VOSTOK
It freed up time for the work that really matters instead of endless spreadsheets and routine. I definitely recommend the course.
Sergey Mills
Producer, MAGIC Production, Star Media
The lessons are built on real, applied tasks, so by the time you finish them you are already using AI agents in your life and projects.
Sergey Butysin
Producer
It was genuinely inspiring. When I tried it myself and got my first results, I thought: “Wow, what just happened?”
Marina Tomarova
Co-founder, MATE Production
A great way into the topic: no overload, plenty of practice. Now I understand how to work with it, how to connect what I need and which part of the routine I can hand to AI agents.
Borya Tomberg
Producer
I took the course for producers and I am incredibly happy with it. Egor explains things calmly and clearly, with great examples and metaphors.
Ira Elshanskaya
Director
Booking
Dates and pricing
€250
per seat · final price
- Includes a group call a week later and 5 video lessons
- In person in Lisbon; the venue is shown next to each date
- The session is not recorded: if you miss it, we move you to another date
- Bring your own laptop
Plus roughly €25 of your own: a ChatGPT subscription and API credit.
- Next session10 October, SatVenue to be confirmedEarly bird: €190 instead of €250 · until 27 SeptemberReserve ↗
- Waitlist24 October, SatVenue to be confirmed
- 31 October, SatVenue to be confirmedEarly bird: €190 instead of €250 · until 27 SeptemberReserve ↗
Seats are paid on Eventbrite. For dates without online booking yet, press the button and write to us: we hold your seat or add you to the waitlist.
Full refund up to 14 days before. Closer to the date, move to another session free of charge if you tell us 48 hours ahead, or send someone in your place. Cancellations and refunds
Do I really not need to code?
You do not. You will connect services, write instructions and set rules — all of it in plain language. Nothing in the 4 hours requires reading or writing code.
Why do I need to spend another €25?
The agent runs on a paid subscription, and connecting your own services costs a few euros of API credit during the session. Both stay yours afterwards. We name it up front so it is not a surprise on the day.
What if I fall behind?
The group is 11 people for exactly this reason. Blocks are paced so everyone finishes each one before the next begins, and the instructor works the room while you build.
Will I be able to do my own task afterwards?
Yes. At the end of the session Egor goes through each participant's personal request, using his own AI assistant trained to design automation plans. You leave with a precise action plan for your task, built on what we covered in the session.
Will there be sessions in English?
Yes. The language is set per date, and the first English session is on 10 October.
4 hours, and it is running
Everyone around you is talking about AI agents. All you need is the time and someone beside you to make your own breakthrough.
Next session: 10 October, Lisbon, in English
Reserve a seat