Find the right problem
We cut through to the workflow that's actually worth changing. Sometimes the honest answer is that AI isn't the solution.
The resultClarity and a concrete next step
Collaboration
Small projects with a concrete outcome.
I take on a small number of focused projects alongside my role at Imerco. Each project has to lead to a clear direction, a prototype or a small tool that works in practice.
Three ways to start
From problem to a concrete next step.
We start with the task, not a particular model or platform. That makes it easier to build the smallest solution and find out whether it's worth taking further.
From task to tool
Understand first. Then build.
A task often starts across emails, spreadsheets and manual checks. I reduce it to one flow, test it, and only build further if it works.
Choose a step. On larger screens, the sequence also follows your scroll.
01 Raw material
The task is scattered.Emails, spreadsheets and manual checks are connected, but no one can see the whole workflow.Email→Spreadsheets→Checks
Build a prototype
I build a focused prototype that can be tested with real people and realistic data before the idea gets expensive.
The resultA working first version
Put it to work
A small tool for one specific workflow, with quality checks and a proper handover. It has to keep working without me.
The resultTool, documentation and handover
A good starting point
You can describe the problem in plain language.
- A workflow takes too long or requires too much manual work.
- An idea needs to be tested before you invest heavily in it.
- A team needs a concrete first example of what AI can do.