The problem we solve
The AI market is fragmented. Advisory firms write reports that end up in drawers. Training bureaus train and leave. Tech companies build tools nobody uses. Almost nobody delivers software that works on the day of delivery and keeps working.
That is why we build, and bring along everything that makes a build succeed: the business case before it, the compliance inside it and the training after it. It is what every build contains, rather than a menu to pick from.
The journey, in order
The start sprint
Week 1
One week, fixed fee. We sit down with the owner of the problem, build a prototype on your data or design the architecture, and put the business case on one page next to it. Then you decide.
More about the start sprint →Business case and scoping
Before the build
What it costs, what it returns, when it pays back. Because we build ourselves, our estimates are realistic. If the business case does not add up, we do not build.
How we make the business case →Build in sprints
Weeks 2 to 12
Short sprints with a demo at the end of each sprint, so you always see where we are. Independently or together with your developers. Connected to your ERP, CRM and the rest of your landscape.
Compliance from the drawing board
During the build
We classify the system under the AI Act, build in logging and human oversight and choose hosting that fits your data.
Training your team
Before the switch
We train the team that will use the system on that system, on the work floor. That also meets Article 4 of the EU AI Act.
How we train →Not tied to one AI vendor
Delivery
We build model-agnostic, so you can switch AI models without rebuilding. Everything comes with documentation, and we keep managing and developing it.
The AI Heroes standard
Included in every build
- 01
Business case first
Before we build, you know what it costs, what it returns and when it pays back.
- 02
Compliance from the drawing board
AI Act classification, logging, human oversight and European hosting are part of the design.
- 03
Your people can work with it
We train the team that will use the system on that system. That also meets Article 4 of the EU AI Act.
- 04
Not tied to one AI vendor
You can switch AI models without rebuilding. Everything comes with documentation.
Compliance from the drawing board
What that means in practice
Classification. Every system we build gets a place under the AI Act: minimal, limited or high risk. That determines what has to be in the design.
Logging and oversight. Every decision the system makes can be traced, and a human can always step in. For an assistant that means handover to a person; for a workflow, a review step.
Hosting. Data that may not leave Europe stays in Europe. On your own servers if it has to be, with a European provider if it can be.
Literacy. The team that works with it is trained on the system. That means you meet Article 4, with the evidence to show for it.
What we hold ourselves to
Working over finished
We measure success by what runs at handover, and by what still runs a year later.
Honesty over sales
If AI is not the solution, we say so. If the business case does not add up, we do not build.
Start small, then scale
First a week, then a decision, then a build. Never a large sum on a hunch.
We do what we advise
Our own assistant runs on this site, hosted in Europe. We automate our own work with what we build for clients.
What it costs
The start sprint costs from €8,000, fixed fee, known up front. A custom build after that starts from €15,000. What your build will cost exactly is on paper at the end of the start sprint, together with what it returns and when it pays back.
Standalone trainings are available too, from €2,500 per half-day; you will find them under training your team. For education and non-profit organisations we apply lower rates.
Next step
Tell us what you want to build
A start sprint is the first step. Below you can tell us what it is about; we respond within 24 hours.
