The pilot worked. Eighteen months later it is still a pilot.
Getting a model to work is the easy part. Getting it into daily operations, across your systems, with someone accountable when it drifts, is the project that actually changes the numbers. That is the one we take on.
What the first call gets you
45 minutes, free, with the architect who leads the build.
- Why the pilot stalled, named concretely rather than diplomatically
- What has to be rebuilt for production and what can be kept
- Which AI Act obligations apply to this use case, if any
- A phased scope with a fixed price per phase, not one big number
€50-150K
Implementation, fixed per phase
8-16 wks
From pilot to production
€778K
Saved per year at a market leader
100%
Yours, code, model and pipeline
You talk directly to a solution architect.
No delivery manager and no discovery invoice. In 45 minutes we go through what has been built, what stalled and what production would actually demand. If the honest answer is that the pilot should be scrapped rather than scaled, you hear that too.
Measured at these companies
Sound familiar?
It works in the notebook, not in the process
A model that needs someone to run it by hand every Monday is not in production. The gap between a working model and a working process is where most of the budget and all of the value sits.
Nobody owns it after the consultants leave
The team that built it is gone, the documentation is a slide deck, and the first time the numbers drift there is nobody who can say whether that is normal. Then it quietly stops being used.
Legal wants to know about the AI Act
For most business use cases the obligations are manageable, but somebody has to determine the risk category, write down what the system does, and keep a log. Left until the end, it blocks go-live.
Curious what this looks like on your data?
Book a free 30-minute call with the solution architect who also builds the work. You walk through your processes together and get an honest answer on where the numbers are for you, and where they are not.
Free intro call
30 minutes, no obligation, directly with the builder.
How it works, without risk on your side
1. Implementation review
45 minutes on what exists, what stalled and why. You get a straight read on whether this is a technical problem, a data problem or an organisational one, because the answer decides the whole plan.
2. Phased, fixed price per phase
Integration, production hardening and handover each priced separately, each with a working result. You can stop after any phase and keep what has been delivered, because you own it.
3. Handover that actually holds
Monitoring, alerting, a runbook, and your own people trained on it. The measure of success is that it keeps running for a year without us, not that we stay on retainer.
Why not a large integrator or your own team?
| Large integrator | In-house team | Stratalytic | |
|---|---|---|---|
| Who does the work | Rotating consultants | People with a day job | The architect who scoped it |
| Time to production | Quarters, with change requests | Whenever there is room | 8-16 weeks, phased |
| After go-live | New statement of work | Depends on one person | Runbook plus training |
| Ownership | Their framework | Yours, if documented | You own code and pipeline |
Before you book
What makes this different from your standard AI project?
Scale and stakes. A first project is one use case in one process, €15,000-€40,000. This is a use case that has to run daily across several systems, survive bad input, be monitored, and be handed to your own team. That is a different discipline, and pretending otherwise is why pilots stall.
We already have a model. Can you take it over?
Usually yes, and often we keep more of it than expected. The model is rarely the problem; the data pipeline, the error handling and the integration are. We review what exists first and say plainly which parts survive and which are cheaper to rebuild than to repair.
What about the EU AI Act?
Most business applications, forecasting, document processing, classification, land in the limited or minimal risk category, where the obligations are transparency and documentation rather than a conformity assessment. We determine the category up front and build the documentation as we go, so it is not a blocker at the end. Where the use case genuinely is high risk, we say so before you commit.
Can our own team maintain it afterwards?
That is the goal. We build on ordinary, boring technology rather than an exotic stack, write the runbook while building rather than afterwards, and train your people on the real system. If your team is not ready to take it, we would rather scope a lighter solution than leave you dependent on us.
Can this be funded?
Often, yes. Production hardening and integration work regularly qualify as R&D under WBSO at 36-50% of the development labour, and MIT can cover 35% of the broader project. We can file both alongside the build and record the hours as the scheme requires.
