syntact
AI pilot · 4–6 weeks · fixed price

Your first AI use case — proven on real data in 4–6 weeks.

For decision-makers who know what they want to automate — but need proof before they commit. Fixed price, written success criteria, a clear exit point.

Discuss a pilot project
Sound familiar?

The use case is clear.
You still don't dare sign off.

The vendor demo was impressive — on their sample data, not on yours.

You're asked to sign an annual contract before anything is proven.

A peer sank €50,000 into an AI pilot that never went to production.

Your IT warns: “our data isn't clean enough for this.”

The department doesn't believe in it — and without them, nothing works.

Nobody can tell you how success or failure would even be measured.

The pilot reverses the burden of proof: measure first, then decide. Not the other way around.

What's included

A pilot that delivers a decision — not a never-ending project.

Built on your real data
Connection to the relevant sources — files, a mailbox, folders, or a system export. The use case runs on your documents, not on sample data.
Written success criteria
Defined before the start: recognition rate, time per case, share without rework. So there's no debate afterwards about whether it worked.
Tested with real users
The department works inside the pilot — not just IT. Adoption shows up in daily work, not in a presentation.
Measured results
Before/after numbers from the pilot run: time saved, error rate, cycle time — your basis for the go-live decision.
Results presentation & go-live plan
The numbers, the open items, and the concrete path to production — presented to your management team.
How the risk stays with us

Criteria not met? The project ends — with no license fees.

Fixed price, locked in before the start — no growing effort estimates
Success criteria defined in writing before anything is built
License fees only start with production use — the pilot commits you to nothing
After the pilot you decide on measured numbers, not slides
FAQ

Before you have to ask.

Yours is missing? Ask us directly.

What does the pilot cost?
A fixed price, locked in before the start — depending on the use case and the data connections required. You know the number before you commit. No effort estimates that grow later.
What data is the pilot tested with?
Your real data — real price lists, real delivery notes, real contracts. That's exactly where canned demos fall apart, and exactly why we don't test with sample data.
Who defines the success criteria?
You do, together with us — before the start, in writing. Typical criteria: recognition rate, processing time per case, share of cases needing no manual rework. Measured against your current reality, not a lab ideal.
What happens after the pilot?
You see the measured numbers and decide: go live or stop. On go-live, the pilot moves to production — system connections, permissions, training. The license only starts with production use.
How much of our team's time does it take?
One subject-matter expert who knows the process, for questions and feedback sessions — a few hours in total over the runtime. Data access is clarified with your IT at the start.
Do we need the AI Quick Check first?
No. The Quick Check helps when you don't yet know which use case comes first. If that's already clear, you start directly with the pilot.

In 6 weeks you'll know whether it works — in black and white.

In the intro call we clarify the use case, the data situation, and the success criteria — and you get the fixed price. Then you decide.

Discuss a pilot project or see the full roadmap first