Assessment
Two or three days on site: looking at processes, checking the state of the data, talking to the people who do the work. The result is an assessed list of possible applications — not a slide deck, but a basis for a decision.
Service area 1
Artificial Intelligence
From a sober assessment of where you stand, through delivery, to running the result. We will also tell you when AI is the wrong answer to your problem.
The real problem is rarely the technology
Almost every first conversation contains the same sentence: “We know we have to do something with AI, but we do not know what.” That is not a knowledge gap you close with another market overview. What is missing is the connection between a general technology and one specific business with grown processes, incomplete data and people who have already lived through one wave of digitalisation.
Making that connection is our work. We look at where time is actually lost in your business, which decisions rest on instinct because the figures are not at hand, and which tasks are uniform enough for a machine to take over. Out of that comes a short list — usually two to four candidates — with an honest estimate of effort, benefit and risk.
Often it turns out that the most worthwhile first step is not AI at all, but a clean data foundation or a tidied-up process. We say so in that case. A project built on unusable data fails reliably — only later and at greater cost.
What we actually offer
The building blocks can be commissioned individually. Nobody has to book the whole path to begin.
Proof of concept
Before a budget is released, we test on your real data whether the idea holds. Kept small and time-boxed. A proof of concept that fails is a good outcome — it has just prevented a project.
Delivery
From prototype to productive solution, embedded in your existing IT. Including connections to current systems, access management and the unglamorous work that turns a demo into an operation.
Ongoing support
AI systems age differently from conventional software: they keep working but degrade quietly as reality shifts. We monitor quality, adjust, and stay reachable.
EU AI Act readiness
Which obligations actually apply to you, and which do not? We classify your systems by risk category, name the concrete requirements and support the work — technically and organisationally.
Enabling your team
So that you do not need us permanently. Training and workshops that work on your own data and tasks rather than textbook examples.
How working together unfolds
Four steps. You can stop after any of them, and after each you hold something that has value even without the next.
-
First conversation
One hour, free of charge, without obligation. You describe the situation, we say whether we can help — and whether it is worth doing at all.
-
Assessment
We come to your business and look at how the work is really done. It ends with an assessed list of possible undertakings.
-
Proof of concept
The most promising candidate is tested against real data, clearly bounded in time and cost.
-
Delivery and operation
The proof becomes a productive solution. After that we support the operation for as long as you want us to.
Training and workshops
So that your team can do it themselves
Buying AI in permanently means staying dependent. A team that has been enabled decides for itself what is worth doing — and spots more quickly when a supplier promises more than they can deliver.
Our training therefore works on your own data and tasks rather than textbook examples. What remains at the end is not a certificate but something you keep: an assessed use case, a prototype, or a rule that applies from the next day.
On site or by video. Groups of up to twelve — beyond that a workshop becomes a conference, and conferences do not produce results.
- Orientation for management Half a day. What AI can do today, what it cannot, and how to recognise a workable use case. No technology required.
- Workshop for departments One to two days. Participants bring their own processes and leave with two or three assessed use cases.
- For development teams Two to three days on your code and your data: integrating models, measuring quality, catching failure cases, preparing for operation.
- Rules for AI use One session, one outcome: a single page setting out which tools are permitted and which data does not belong in them. The most common blind spot in mid-sized companies.
The cheapest way in costs five minutes.
The AI Check gives you a first assessment without having to speak to anyone. Afterwards you decide whether a conversation is worth it.