01The Problem
AI adoption usually fails the same way: a company buys licences, sends a memo, and six months later the tools sit unused, or get used badly. The technology isn't the problem. What's missing is the translation between what AI can do and the actual tasks on someone's desk.
Non-technical people don't need to understand transformers. They need to know which of their daily tasks AI can absorb, how to hand those tasks over safely, and how to check the work.
02Teaching at Scale: State-Sponsored Bootcamps
Through coop.tech, the first Portuguese tech cooperative, which I co-founded and preside over, we built an AI educational tool for teaching programming. It has been used in state-sponsored bootcamps, with students coming into tech from very different backgrounds.
One lesson from the classroom applies everywhere: people learn AI by using it on their own problems.
The work is also informed by EU AI governance. CTVC holds a seat on the EU AI Act Advisory Forum, where it provides technical opinions to the European Commission, and it is one of only two cooperatives in the Forum. We know the regulation from the inside.
03Inside Companies
In the companies I run, we treat AI adoption as an engineering problem. The approach:
- Start from the task. Map what a person does in a week, then find the pieces AI can take over: reporting, documentation, summaries, first drafts, cross-checking.
- Build the workflow around them. A non-technical person shouldn't have to configure anything. They work through tools they already use (chat, email, forms) and the AI plumbing lives behind that.
- Add guardrails. AI output gets checked where it matters and flows freely where it doesn't. People stay accountable for the results while AI absorbs the mechanical work.
- Train by doing. Short sessions on real tasks work better than long courses on theory.
The measure of success: we don't count how many people have AI licences. We count how many tasks stopped being done by hand, and whether that is still true a month later.
04Try It: Pick a Task
Pick a task any team does weekly, and see how the workflow changes. The examples are made up. The real version of this exercise is done on your team's actual tasks.
05Where This Goes
The same playbook (task mapping, invisible plumbing, guardrails, training by doing) transfers to any team that pushes paper, writes reports, or answers the same questions repeatedly. The technology is ready. The bottleneck is the translation, and that is the part I do.
Want your team using AI instead of just paying for it?
I run enablement programs built around your real workflows.