There's a moment in every automation project that no one warns you about. The moment when the system works perfectly, the indicators are green, the agent does exactly what it's supposed to — and the team is still doing it by hand.

Not out of stubbornness. Out of unfamiliarity. Out of distrust. Out of lack of a ritual that integrates the new way of working into the actual routine.

The invisible design error

When we build a system, we think about the flow, the logic, the use cases. We rarely think about the first time a real person will have to interact with it without anyone by their side.

We've seen excellent systems that never took hold because there was no clear "how to use this". Not a 40-page manual — a 20-minute process and a use context.

What AI can't automate

AI can automate a task, but it can't create the trust that makes a team delegate to the system, or convince someone that their way of working has to change, or detect the colleague who does things differently and silently breaks the circuit.

That's human work. And it doesn't happen by itself.

What we've learned

Projects that work don't end when the system is delivered. They end when the team has integrated the new way of working as a habit. Two weeks of active follow-up are worth more than six months of documentation.

It's not an extra step. It's half the work.

If you have a system that nobody uses, let's talk.

→ Related article: Why the first month is usually the hardest