Published 11 June 2026
AI success stories tend to feature large companies with their own data teams. Reality in an SME is different: no data scientists, no experimentation budget, plenty of daily pressure. The good news: that is exactly where the return per invested euro is highest. At Medux, one well-chosen AI application delivered a 70% cost reduction on a labour-intensive process.
The step-by-step plan
Step 1: pick one process, not a strategy
The classic mistake is starting with an organisation-wide AI vision. Start the other way around: pick one process that demonstrably costs time. Writing proposals, processing inbound email, drafting reports. One process means a manageable experiment and a measurable result.
Step 2: measure the baseline
How many hours does the process cost per week, and what goes wrong? Without a baseline you cannot prove what AI delivers later, and those numbers are exactly what you need for the decision to scale.
Step 3: train the team first
The cheapest AI investment with the highest return. A team that understands what AI can do and where it fails spots opportunities in its own work and uses tools safely. A one-day in-company training lays that foundation, and covers much of the AI literacy obligation from the AI Act at the same time.
Step 4: try existing tools before building
Many SME processes improve with tools that already exist: ChatGPT or Claude with good working instructions, Copilot in Microsoft 365, AI features in your current software. Custom building only makes sense once existing tools demonstrably fall short.
Step 5: build small, prove fast
If custom work is needed, demand a working prototype within weeks, not a months-long programme. We build proofs of concept in days (at Trabu: idea to working product in 6 days) so you decide on the follow-up investment based on something real.
Step 6: scale what works, stop what does not
Compare against your baseline from step 2. If it works, roll it out to the next process. If it does not, stop without losing face: a stopped experiment of a few thousand euros is not a failure, a pushed-through failure of a hundred thousand is.
The three pitfalls
- Starting with technology instead of the process. The question is not “what can we do with AI” but “where is time leaking and can AI plug it”.
- Skipping adoption. A tool the team does not trust or understand dies quietly. Training and involvement from day one.
- Forgetting compliance. The first AI Act obligations have applied to SMEs since 2025. Arrange the basics now; see our AI Act guide.
Help with the first step
If you are unsure where to start, the AI readiness assessment (1 to 3 days, from EUR 3,000) is the shortest route to a concrete answer: where your organisation stands, where the opportunities are and what the logical first step is.
Frequently asked questions
Is AI too expensive for SMEs?
No. The entry point is a training (from EUR 2,500) or a readiness assessment (from EUR 3,000). Most AI gains in SMEs come from using existing tools smarter, which costs attention rather than capital.
Where does AI deliver the most in an SME?
In processes with lots of repetitive reading and writing: proposals, customer questions, reports, administration. Rule of thumb: wherever an employee spends more than an hour a day on documents or email, there is an AI opportunity.
Does our SME need to comply with the EU AI Act?
Yes, the AI Act has no SME exemption. The AI literacy obligation has applied to everyone working with AI since February 2025; Dutch enforcement starts 2 August 2026.