What we believe about building with AI.
These are not slogans. They are the rules our team, our tools and our AI assistants already work by every day — written down so clients know what to expect from us, and so we can be held to them.
Ten rules we work by
Truth before speed
An assistant — or a person — says “done”, “sent” or “created” only when a system confirms it. When a record is empty or out of date, we say so.
AI drafts, humans send
Nothing leaves for a client, a prospect or the press without a person deciding to send it. Our models prepare the work; people sign it.
Machines never delete
We archive, stop and rename. A deletion is a human decision, confirmed and logged, and there is always a copy until it is.
Every action leaves a trace
Decisions, requests, blockers and what was actually done are recorded where the whole team can find them.
Depth in one industry beats breadth
Our AI Pilots are built around the workflows of one sector at a time, because that is where AI earns its place.
We run what we build
Each project is a container stack on infrastructure we operate, with backups we have actually restored. See how we build.
No secret in code, ever
Keys live on the server, set through a controlled path, and are never read back in production. A key that touched git is treated as leaked.
Real data stays real
Demos and previews run on fake or anonymised data. Personal data is handled only where it must be, by named people.
Everyone can ship
Anyone in the team can start a project and see it live in minutes. What can hurt — money, domains, deletions, personal data — has one clear owner.
Useful deliverables only
A proposal, a quote, a working product: yes. A document nobody will read or hand over: no.
Why write it down
AI makes it cheap to produce words, code and promises. What stays expensive is trust: knowing that a system did what it says, that nothing was lost, and that a person answers for what goes out. These rules are how we keep that trust — in Takat OS, in the agents we build for clients and in how we talk about our work.
Work with a team that says what is true.
Tell us what you want to build. We will tell you what we would do, and what we would not.