Executive program by Jorge Fraga · jorgefraga.ai
Using Copilot is easy. Operating through it is a skill. This program teaches the second one: choosing the right surface for each moment of work, giving it the right context, and validating the result before acting. Copilot stops being a feature and becomes an AI-assisted work layer applied to communication, analysis, documentation, meetings, decisions, and automation, always in an enterprise context and with governance built in.
By the end, every participant has a real task worked from the first request to the verified result, a team assistant built and tested, routines that run on their own with a human owner, and a method that survives the next renaming of the tools.
<aside>
💡
📊 OVERVIEW
- Who it is for: managers and business leaders who decide on the basis of scattered information; technical and operations teams who live in data, tickets, and documentation. The program has a variant for each profile, with the same modules and different cases.
- Context: organizations in the Microsoft 365 ecosystem that want to use AI with control, practical utility, and alignment with enterprise and regulatory requirements.
- Format: four modules, 24 hours (6 + 8 + 5 + 5), adaptable between 8 and 24 hours depending on the audience. In person or online. Closed cohort, internal academy, or executive education.
- Positioning: from the first prompt to a tested team assistant, with human validation at every step.
</aside>
🧩 THE METHOD
Five models structure the program and run through the four modules. The tools change every quarter; these do not.
- Six elements plus limits. Persona, instruction, context, format, audience, and tone, plus what cannot happen. A prompt is a work instruction, not a sentence.
- Six quality criteria. Correctness, coverage, fit, format, traceability, and tone. They evaluate any answer, and Copilot can be asked to evaluate its own against them.
- Four steps of analysis. Understand, build, communicate, challenge. Always in the same order, in Excel and beyond.
- Four levels of autonomy. Prompt, saved prompt, agent, scheduled prompt. What changes between levels is who decides when the work starts.
- Seven habits of the augmented operator. Understand before you change, say which surface you are in, give the format and not just the request, ask what the analysis could be wrong about, promote what works, schedule the recurring with a human owner, sign the output.
The rule that binds the five: AI prepares, the person validates and decides.
📚 PROGRAM STRUCTURE
Module 1: Foundations, Safe Use, and Prompt Engineering (6 hours)