Plain steps. No theatre.
Every project follows the same path, whether it's a four-week AI pilot or a six-month platform build.
Brief
You send a short brief or book a call. We ask the questions that change the price: inputs, outputs, environment, deadline, what already exists.
Proposal
A written proposal with scope, milestones and a fixed price for each. For AI work, it includes the success criteria the pilot has to meet.
Build in milestones
Each milestone ends with something you can run or hold. You review it, we adjust, then we move on. Scope changes are agreed in writing, not discovered on the invoice.
Test on the bench
Hardware is tested on our bench with a setup like yours; software is tested against your real data. You get videos, logs and results, not just a "done".
Handover and support
Source, schematics and documentation are yours. We stay until it works on your side, and we're here when you need the next version.
Four weeks to a real answer
The safest way to start with AI: small, fixed-price, on your own data, with a decision at the end.
One goal, real data
One measurable goal, access to your real data or equipment, and success criteria in writing.
A working model, wired in
A model on your data, connected to one real input and one real output.
Results and a decision
Live demo, results against the criteria, and a go or no-go with a costed production plan.
Frequently asked
What does a project cost?
It depends on scope, so we quote each one. Most projects are fixed-price with milestones, agreed after a short scoping conversation. Small or open-ended work can run hourly. The AI pilot is a fixed price agreed before we start.
Do you sign NDAs?
Yes. Tell us in your brief and we'll sign your NDA, or send ours, before you share any details.
Who owns the code and designs?
You do. Source code, schematics, PCB files and documentation are handed over when the project is paid.
Can you work on hardware remotely?
Yes, it's most of what we do. We build and test on our own bench with a setup as close to yours as possible, send you test videos, and ship the finished unit. When something behaves differently on your side, we troubleshoot it with you.
Which AI models do you use?
Whatever fits the job and your rules on data: models from OpenAI, Anthropic and Google, or open-weight models we can run on your own hardware so data never leaves your site.
What if AI isn't the right answer?
Then we'll say so. Plenty of problems are solved better by a sensor, a rule or a database query. The pilot exists so neither of us bets big on an assumption.
Can you work with our existing code or equipment?
Yes. A lot of our work is extending, fixing or connecting things that already exist: legacy C and C++, LabVIEW projects, PLCs, ERP systems and boards someone else started.
How do we communicate?
One point of contact, regular written updates, and calls in your time zone when they help. We work remotely with clients in 28 countries.
Tell us what you're building.
Send a short brief. We reply with questions or a written proposal with a fixed price and milestones. If AI isn't the right answer, we'll say so.