AI decision workshops for European teams
Your team already knows AI could help. This is the day they work out exactly where and how.
Half a day, one real workflow. Your team decides what AI should take on and what stays human, we pick the tools that fit, and we run it together until it produces something you'd actually use, in a fraction of the time it takes today.
Tell us about one workflow your team dreads →For professional services, creative studios, operations and other teams doing hands-on work.
AI rollouts stall because of the room, not the people in it.
It isn't the tools.
We interviewed leaders running AI transformations at European startups and scale-ups. The pattern wasn't that the tools didn't work. It was that adoption comes down to new habits, and nothing around the work had changed enough to build them.
Adoption holds when the team decides.
When a team decides together which parts of its work AI should touch, and its leader stands behind it, the change lasts. Left to individuals, it drifts. That decision, made together and backed from the top, is what we facilitate.
That's a room problem, not a technology problem. It's fixable in an afternoon, and that's what this is.
One real workflow. Five moves.
We start with one workflow because that is how the method sticks. Your team learns it by running it on real work, then takes it to the next one.
Map the work
We start with how the work actually happens today: one real workflow, step by step. Not a hypothetical use case.
Sort the ladder
The team places every step on a rung: keep human, augment, delegate, or automate. The people who own the work make the call.
Pick the tools
We are tool-agnostic. ChatGPT, Claude, Gemini, Notion, or whatever you already pay for: we choose per step, based on the work, not on who we resell.
Run it together
We rebuild the workflow in the room and put it through real work, with your team at the keyboard, until the output is good enough to use.
Keep the kit
The task cards, the workflow map, and the session format stay with your team. The next workflow shouldn't need us.
The agency ladder
The agency ladder is a four-rung scale for sorting every step of a workflow: keep human, augment, delegate, automate. Agency as in human agency, meaning who stays in control of each step.
AI doesn't touch this. The judgment, the relationships, the parts that make the work yours.
You still do the work — AI makes it faster or sharper.
AI does the work — a person reviews and owns the result.
AI runs it end to end, with a name attached and a way to check.
Every step of the workflow lands on one rung — chosen by the people who do the work, not by us.
Two ways in.
We work with small and mid-sized European companies: the ones the big firms overcharge and the free programs underserve.
Decision workshop
Half a day on one real workflow. Your team sorts every step on the ladder, picks the tools, and rebuilds the process with us in the room. You leave with the workflow running, the parts you deliberately kept human, and the kit to do the next one yourselves.
Adoption program
We help the decisions hold. Over 90 days the leader keeps backing them, the person carrying them inside your team gets our support, and we check in on every experiment while it runs. You end with a before-and-after on one workflow: a baseline on day one, the same measure at day ninety, and whether people were still using it at week six. Alongside the numbers, an honest read on how the change landed, because a workflow that got faster and made everyone miserable hasn't worked.
Start with the workshop. The program makes the decisions stick. Both are a fixed fee, quoted before you commit.
If something isn't being adopted, we treat that as a signal to fix it or drop it — not to push harder.
Both end the same way: the kit stays with your team.
What this isn't.
They build it for you and hand it over. We build it with your team, in the room, so they can run the next one without us.
They teach AI in the abstract. We use it on your actual work, on the day, and leave the method behind.
Buying the tool was the easy part. The using is the work.
The gap isn't skills. It sits one level up, in the team and the conditions its leader sets.
AI that erodes trust fails. AI that elevates human contribution sticks, and pays.
"No" is a real answer. A team deciding some work stays human isn't resistance — it's the method working.
Who you'd be working with.
Ottavio Cambieri
Founder · Tallinn, Estonia
Founder of Naked Oracle, a product and marketing consultancy in AI, tech, and the creative industries since 2019, and guest lecturer in interaction design at the Estonian Academy of Arts since 2018.
People1 AI grew out of his research at the same academy into how AI adoption reshapes work: what happens to people's sense of their own work when tasks get handed to AI. The method on this page came out of it.
Based in Tallinn, Estonia. Works across Europe, remote or on-site. Sessions run in English or Italian.
Ottavio on LinkedIn ↗What to expect.
The people who own the workflow, and their leader. The decisions only hold when the leader is there to back them, so that part isn't optional.
A team, not an audience. If everyone can argue over one workflow at one table, it's the right size.
Remote or on-site, wherever your team works. Same method either way.
Pick one workflow. That's the whole prep, and the form below already starts it.
The task cards you sorted with, the map of your workflow, and the session format. Yours to keep.
What this gives you under the AI Act
Since 2 August 2026, national authorities can enforce Article 4 of the EU AI Act: any organisation deploying AI must ensure its people have sufficient AI literacy for their role. There is no exemption for small companies.
There is also no certificate to buy. The Act sets no curriculum and no exam — it expects measures proportionate to your context, and evidence that you took them.
The workshop produces that evidence as a by-product: a mapped workflow, a documented decision for every step about what stays human, a dated participation record, and the materials your team keeps. That is a contemporaneous record of a deliberate, role-appropriate measure.
It is not certification, and nobody can sell you certification for this. If someone offers it, that's a reason to be careful.
Tell us about one workflow your team dreads.
A few lines is enough. We'll reply with an honest read — including whether AI should touch it at all.
Prefer to talk it through? Book 20 minutes ↗