Let’s talk Three two-hour sessions to turn the task that eats most of your time into a process that AI sustains. On your own company, with no coding and no tool catalogs.
For owners, managers and founders who decide on budget, team and processes.
Sales proposals
What does it cost today?
45 min × 3/wk × 2 people ≈ 19 h/month
What information does it need?
Client brief, rate card, past proposals
Where is the judgment?
Scope and discounts: a person decides
What if it goes wrong?
Client commitment → mandatory review
How do I know it's right?
Six-point checklist before sending
October 14, 21 and 28
18:00 – 20:00 · Lima time · Remote, live
It's the first question of the method. Put a number on the task that takes most of your time and decide whether it's worth touching. The formula is the same one you'll use in the course: minutes per run × frequency × people.
How long one full execution takes, start to finish.
People involved
Current cost of the task
this task
9.7
hours a month
A reasonable candidate
Between 4 and 10 hours a month. Worth turning into an assisted process if the information it needs is within reach.
It goes straight into your diagnostic questionnaire.
This measures the cost of the problem, not the savings. How much you recover is calculated at the end of the course, with your process already running.
It's the time you never get back by hiring. That's why we deliberately leave out what almost every other offer includes.
Our edge is depth and transfer, not coverage. Anyone arriving for breadth will be disappointed in the first session, and we'd rather say it here.
Your process enters in Phase 0 and stays with you to the end. Reference cases are used to show the method, never as a template to copy.
You declare your most repetitive tasks and get accounts and access ready. It's the enrollment form on this page.
By the end
You arrive at session 1 with your candidates already declared.
Between sessions: An active account on one platform, a laptop and access to your email and files.
What a model does well and badly. Instructions with goal, context, rules and format. Persistent context spaces, files and research with sources.
By the end
You pick the process you'll work on for the rest of the course and put a cost on it.
Between sessions: Fill in your process sheet with the real information it needs. About an hour.
The five-question sheet. What gets delegated and what stays under your judgment. Building the assistant live, with real information from your company.
By the end
Your assistant created and running on your information, even if it still fails.
Between sessions: Run the process three times in real conditions and log where it fails. About an hour.
Fixing your real failures. Email, file and calendar connectors. Agents and scheduled tasks with an approval point. Measurement, confidentiality and what to buy.
By the end
The process running with validation and measurement defined, plus the map of the next ones.
Between sessions: 30 days of follow-up questions and a review session to measure the sustained result.
It's the core mental model of the course: break any process into input, context, rules, analysis, validation and output, and answer five questions on a single sheet. The answers decide how much gets delegated. Not enthusiasm for the tool.
Minutes × frequency × people. Decides whether it's worth touching.
Where it lives, who has it, in what format. It's the context AI needs.
Separates the transferable part from the part you don't delegate.
Determines how much autonomy the system is allowed.
Validation is designed before building, not after failing.
Climbing isn't always worth it. The course takes you to level 3 and shows you when level 4 makes sense.
A person does practically everything.
AI speeds up one specific part of the task.
Context, rules, steps and validation defined.
AI executes or triggers actions; you step in where your judgment matters.
AI brings capacity, not judgment. A process that's 70% assisted and 30% controlled is a good result, and it's presented as such.
None of this is a prompt to copy. Every piece forces you to think through your own process, which is what makes the method work for the next one without an instructor.
A task you used to do by hand that now runs with AI in a structured way, with validation defined.
A project, workspace or assistant configured with real information about your role or business.
Time per run before and after, and hours freed per month, to defend in front of your team or your board.
Three to five additional processes evaluated and prioritized by cost.
Every request that comes through the website produces a quote drafted by AI on our real rate card. It's born as a draft and an advisor reviews it before sending. It's the sales-proposal case shown in session 2, and you can try it now.
Try the quoterThe price includes the prior diagnostic, the three sessions, the full kit and the follow-up. No separate modules.
Diagnostic due by October 2, 2026
Seats are confirmed in order of enrollment, once the diagnostic is answered.
Price in US dollars. If you're in Peru, we invoice in soles at the day's exchange rate.
Ten minutes. You declare the tasks that take most of your time, and we use them to prepare the first session's demos on real processes from the people enrolled.
Start with the heaviest. One is enough to enroll; with three you arrive at session 1 with options.
No. Code shows up only as something the AI itself produces and you review. The whole course runs in business language: inputs, rules, owners, validations and control.
A paid ChatGPT, Claude or Gemini account. Several session-3 features, like connectors and scheduled tasks, don't exist on free plans. The confirmation email tells you which plan to get and what each one enables.
No, and it's a deliberate decision. It's the best-selling topic on the market and almost every offer includes it. This course works on the owner's administrative and management load, which is where the time you never get back by hiring lives.
The calculator on this page and the enrollment questionnaire exist for that. If you arrive without clear candidates, session 1 has a short catalog of typical tasks by area to choose from on the spot.
The current edition is remote and live. You build your assistant on your own laptop, with your information, while the instructor guides each participant.
You get 30 days of follow-up questions about the process you built and a review session at the end to measure the sustained result and unlock the next one. Without follow-up, some processes get abandoned in the first week, and we don't want that.
Yes, as long as they have decision power over the process they'll work on. Each participant builds on their own task, so two people from the same company can bring two different processes.
The group is small on purpose. If no seats are left, the same form puts you on the waitlist and we let you know first when one opens or we add a date.