From one process to a system that runs.
We start narrow on purpose: a narrow thing that works beats a broad thing that nearly does. Then we extend it, one proven step at a time.
The same process, from the way it runs today to the way it runs itself.
- Stage 01: Discover
Map the process as it really runs.
We sit with the people who do the work and follow it through your tools (the re-keying, the chasing, the checking) to find where AI pays back first, and what to leave alone.
- You get
- A written map of the process, and what we would automate first
- We need from you
- One decision-maker and an hour with the people who run the process
- Stage 02: Design
Write it down before we build it.
The workflow, the rules, the agents, the approvals and the limits, agreed in writing. You see exactly what will be built, where a person decides and what it will cost.
- You get
- A written scope and a fixed price
- We need from you
- Your policies, your edge cases and a yes or no
- Stage 03: Build
Build it on the tools you already use.
We connect the systems, write the rules in code and build the agents, each with only the tools it needs and a budget it cannot exceed.
- You get
- Agents and workflows connected to your systems
- We need from you
- Access to the tools involved, scoped to what the work needs
- Stage 04: Validate
Test it on real examples.
Before anything goes live it runs on your real documents and cases, and every output is checked against what your team would have done. What it cannot do is refused by name, not guessed.
- You get
- A tested system, and a record of every case it was checked on
- We need from you
- A set of real examples, and someone to judge the results
- Stage 05: Deploy
Go live under supervision.
The system goes into production with its gates in place: anything that leaves your business starts as a draft or a proposal a person approves.
- You get
- A live system, with approvals where they matter
- We need from you
- The people who will approve, and a date
- Stage 06: Operate
Monitor it, maintain it, improve it.
Once live, we watch every run and every cost, keep each connection working as your tools change, and improve the system month by month, extending autonomy only where its record has earned it.
- You get
- A system that keeps running, and keeps getting better
- We need from you
- Your feedback, and the next process when you are ready
What stays true at every stage
- 01
One process first
We start with the one process that pays back soonest, and prove it before extending.
- 02
Written before built
Scope, rules, approvals and price are agreed in writing before any work begins.
- 03
Refused, not guessed
When something cannot be automated where it stands, we say so before you pay for it.
- 04
Autonomy is earned
Every system starts supervised, and takes on more only as its record shows it can.
What is yours
What you own, what you keep if you leave and in what form is written into your proposal before anything starts. The platform the work runs on is ours: it is what makes the second workflow cost less than the first.
The questions operators ask first.
How fast can we start?
As soon as you book: the first call is free and takes thirty minutes. A focused automation can be live within weeks; larger systems and departments are delivered in phases, each one working before the next begins. Your written scope says when each part goes live.
Who owns what you build?
It is written into your proposal before anything starts: what is yours, what you keep if you leave, and in what form. The platform the work runs on is ours: it is what makes the second workflow cost less than the first.
What is AI Operations, and do we need it?
It is how a live system stays healthy: we monitor every run and cost, keep connections working as your tools change, fix what breaks and improve the system every month. Anything in production should have it; the plan scales with the size of what you run.
How much does it cost?
Builds start from €750 for a focused automation, €2,500 for an AI agent, €5,000 for an AI system and €10,000 for an AI department; enterprise programmes are priced per scope. Keeping a system monitored and improving starts from €250 a month. The real figure depends on scope, integrations, data, rules and support, so after a free assessment call you get a written scope and a fixed price before anything starts.
Start with the process you'd fix first.
Thirty minutes, no pitch deck. We walk through how the work runs today, show you the platform running on a demo company, and leave you with a written summary of what we would automate first.
Free, 30 minutes, and a reply from a person within one working day.