Published 11 June 2026
Search for AI implementation costs and you will mostly find “contact us for a quote”. We simply publish our prices. These are the price bands we use, and with some variation they reflect the Dutch and European market.
The prices in one table
| Service | Price band | Timeline |
|---|---|---|
| AI training (in-company, per day) | EUR 2,500 - 4,000 | ½ - 1 day |
| AI readiness assessment | from EUR 3,000 | 1 - 3 days |
| Consulting (roadmap, business case, guidance) | EUR 3,000 - 50,000+ | weeks - months |
| Proof of concept | from EUR 15,000 | days - weeks |
| Custom AI software | EUR 15,000 - 100,000+ | weeks - months |
Where the money goes
Training: EUR 2,500 to 4,000 per day
In-company AI training is the cheapest investment with the broadest effect: your whole team works faster and more safely with AI afterwards. The price varies with group size and customisation. Compare that with the hidden cost of not training: employees reinventing every wheel, or pasting company data into public tools.
Consulting: EUR 3,000 to 50,000 and beyond
At the lower end sits the AI readiness assessment (from EUR 3,000): in 1 to 3 days you know where you stand and what the logical next step is. A full AI roadmap with quantified business cases or long-running implementation guidance sits in the tens of thousands. Rule of thumb: the bigger the decision that follows, the more the groundwork may cost.
Software: EUR 15,000 to 100,000 and beyond
Custom AI software has the widest spread. A well-scoped application, like automating one document flow, sits at the lower end. Systems with multiple integrations, high accuracy requirements or heavy compliance sit at the top. We almost always start with a proof of concept: a working prototype in days (at Trabu, 6 days) before committing to the full build.
The costs nobody budgets
Three items that are often missing from quotes but decide the business case:
- Data preparation. If your data is scattered, polluted or unstructured, time goes there before AI can do anything with it.
- Adoption. A working system nobody uses returns nothing. Reserve budget for training and change management.
- Maintenance. Models, integrations and regulation keep changing. Count on an annual amount for upkeep and iteration.
What does it return?
Costs without returns mean nothing. At Medux, a custom AI solution delivered a 70% cost reduction on a labour-intensive process. Not every project hits numbers like that; that is why we quantify the business case before building, and sometimes advise not to build at all.
Want to know what AI implementation would cost your organisation? Start with a readiness assessment or plan a no-strings conversation.
Frequently asked questions
What does it cost to get started with AI?
Getting started does not have to be expensive. An AI readiness assessment costs from EUR 3,000 and a training day for your team from EUR 2,500. That tells you where the opportunities are before you make bigger investments.
What does custom AI software cost?
Custom AI solutions run from EUR 15,000 for a focused application to EUR 100,000 and more for complex systems. A proof of concept to validate the idea sits at the lower end of that band.
What drives the price of an AI project?
The biggest price factors are the complexity of your data, the number of systems to integrate with, the accuracy and compliance requirements, and how much custom work is needed on top of existing models.
Does AI implementation pay for itself?
With a well-chosen application it does. From our own practice: at Medux an AI solution delivered a 70% cost reduction on a labour-intensive process. Payback time depends on the application; always run the numbers before you build.