$ masli services --priced-up-front
Three ways to work with us.
The first two carry prices because you should be able to size the decision without booking a call. The third does not, and the reason is below rather than hidden.
-
01
Consulting & roadmap
from €2,000
We work through your processes with the people who run them. What comes out is a written roadmap: where AI changes the outcome, in what order to do it, and what each step is worth.
- A map of the actual workflow, not the org chart
- Each opportunity ranked, with the value attached
- A written roadmap your own team can execute
- Acceptance criteria written for whatever you build next
-
02 · most common
Adapted solutions
€5,000–20,000
Building blocks that already work, fitted to your processes, your data and your stack, and left running in production.
- Acceptance criteria agreed in numbers before the build starts
- Model-agnostic — no lock-in to one vendor’s API
- Runs in your accounts, in your repository
- Handover includes the checks that keep confirming it works
-
03
Custom build
Scoped individually
When nothing off the shelf fits. Scope, timeline and price are agreed in conversation, after the problem is understood.
- Scoped against a written acceptance condition, like the others
- Fixed scope per phase, not an open-ended arrangement
- Everything handed over as plain files in your repository
In every engagement
Four things do not change with the price.
The size of the engagement decides how much gets built. It does not decide whether the work can be checked, or whether you own the result.
- Success is written as a measurable condition before work starts
- always
- Deliverables are plain files in a repository you own
- always
- Model and vendor choices are made reversible
- always
- Hosted services of ours you would depend on afterwards
- none
Questions we get asked first
Why is the third tier not priced?
Because any figure we published would be wrong for almost everyone who read it. Work at that size is scoped after the problem is understood, and the number falls out of the scope rather than the other way round. Publishing one anyway would only filter people out on a guess.
What if the roadmap says we should not build anything?
Then that is the finding, and it is the cheapest one you will ever buy. The engagement is priced so that this outcome is not a loss for you, and the roadmap is deliberately written as something you can act on without us.
Do we need to have chosen a model or a vendor first?
No, and we would rather you had not. Which model runs behind a step is an implementation detail, and we build so that it can be swapped. A system that cannot survive its vendor changing terms is not finished.
Who owns what you build?
You do. The schema, the code and the checks are plain files in your own repository — no proprietary format, no hosted service of ours in the middle, no dependency on our being reachable. If the arrangement ends, what was built keeps working.
Do you work with our team, or instead of it?
With. The part that decides whether an AI system is still working in six months is whether somebody can tell that it has stopped — so the checks, and the reasoning behind them, are handed over alongside the system itself.
Can we see the method before committing to anything?
Yes. On a call we run the machinery live — the integrity checker over a real corpus, the commit history, the structural diffs — rather than showing slides about it. It either passes in front of you or it does not, which is the point of doing it that way.