Training
It is easier to use AI badly than well
Badly looks fine: fast, confident answers that read like a robot wrote them. The Cook-a-Long teaches you to use it well, cookery-class format: a real job, everyone building along live, and a working tool by the end. No slides read at you.
Free, weekly and open to anyone. This exists because we would rather teach you than bill you: the whole philosophy is at do it yourself.
generic corporate speak, sounds like every other AI email
You write follow-up emails for our sales team.
Existing customer, call this morning about the new product.
Draft the follow-up email from my call notes.
Three short paragraphs, one clear next step.
My tone, under 120 words, no hype.
same AI, completely different email
Real counts from the sessions, not marketing maths. The tools are things like follow-up email drafters, research assistants and meeting-prep Gems, built by the people who now use them.
Why this exists
Once you know, you know
I taught myself everything Operosus runs on. Before AI, that meant years of being blocked: waiting on developers, queueing behind other people’s priorities, paying through the nose for software that was nearly right.
AI removed those blockers for me. It can remove them for you, but there is a catch nobody puts on the box: the gap between using it badly and using it well is enormous, and badly is the default. A few habits close that gap. They are teachable in an hour, and once you have them you cannot go back.
Education is the way in. Everything else we sell is optional.
Inside a session
This is what an hour looks like
The Module 4 build from the session format: an attendee goes from a blank Gem to a working follow-up email drafter, tested against their own call notes, before the hour is up.
“Superb session, great insight and fantastic to leave with some genuine oven-ready takeaways.”
The format
The session, module by module
Two frameworks carry the whole session, then four modules put them to work. Every prompt below is a real one from the session, you build with it live.
framework 1
The 4 Cs
Four things that improve any prompt you write. Vague requests get vague results.
framework 2
Prompt anatomy
Five building blocks that structure a strong prompt. The hero animation above is this framework doing its job.
AI ideas generation
Use AI to work out how to use AI. It has seen thousands of use cases, so it spots where automation makes sense faster than you can. You pick the ideas worth trying.
> Give me 8 specific ways someone in my role, with my responsibilities and challenges, could use AI to work more efficiently.
Smart writing
Why most AI writing sounds like a robot, and the fix. The weak prompt, the rewrite, then a template with blanks you fill in yourself so the habit sticks.
> Show me that weak prompt rewritten with Role + Context + Task + Format + Constraints.
Research and analysis
Hours of desk research becomes a sourced briefing. Deep Research scans hundreds of sources; NotebookLM answers questions from your own documents, with citations on everything.
> Create a step-by-step Deep Research prompt for this question. Keep the output to a maximum of 3 pages.
Personal AI assistants
Build a Gem for one task you repeat every week, test it on real work, keep it. Set it up once and reuse it, instead of prompting from scratch every time.
> Give me the exact instructions to create this Gem, structured with Role + Context + Task + Format + Constraints.
Try the free weekly session
Open to anyone, no pitch, no follow-up sequence. Turn up, build something, keep it.
Join the next Cook-a-LongRun it for your whole team
Private Cook-a-Longs built around your tools, your data and the jobs your team does every week. Everyone leaves with something running.
Book private training