Fifteen years in enterprise IT. Now we build with AI.

Most of IT started on the service desk and in the field. From there it went to leading engineering work, running IT as an interim manager, and leading workplace services as a product owner for organisations with thousands of people. Identity, devices, access and security: the parts that have to work every morning and can never leak.

That is still the base. What changed is what we build on it. We put AI into the daily work of growing companies, with the care a large organisation would demand, and we teach teams to think AI first.

David Omisi standing in front of a large sculpture in Rotterdam
David Omisi, founder. Your first conversation is with him.

Large-company habits, at your size

We know what breaks
Much of our work has been taking over environments that grew without a plan and leaving them clean and governed. We build for the day after go-live, because that's where most projects fail.
Enterprise standards, at your size
Named accounts, the least access that works, logging, and a written runbook. The same rules we applied for organisations with thousands of people, without the paperwork that usually comes with them.
Made to grow
What we build for one team runs in your accounts, on tools you already pay for, and is written down. When the next department wants it, you extend it instead of starting again.
Remote by default, onsite anywhere
We work for clients in different countries, so most work happens remotely. When the job is better done in the room, we come over, at home or abroad.

We work AI first ourselves

Before we ask a team to change how it works, we change our own. This is part of what that produced in the past year.

How we train teams to work this way

  • Connectors that let an AI assistant read and act in Microsoft 365, under its own limited account.
  • Teams of AI agents with fixed roles, where one model's work is reviewed by another before it counts.
  • Data collectors for systems that have no API at all.
  • Bookkeeping flows that read bank and payment feeds and prepare the entries.
  • Routing that sends each task to the model that does it best, across several providers.
  • A software product, from first sketch to live.

What shouldn't your team be doing anymore?

Tell us what it is and who does it now. If it's a fit, the next step is a 30 minute call.