FAQ

We take boring, recurring work off your plate and make it run itself. We build the loop: deterministic code does the heavy lifting, AI handles only the narrow parts code can't, and checks and fail-safes keep it honest. The work still gets done, and it gets done faster, sharper, and more reliably.

You can open a chatbot and prompt it. But when you're prompting, correcting, and re-prompting, you're working for it as much as it's working for you. You've become the system: the memory between steps, the checks, the one catching its mistakes. That's fine for a one-off. It falls apart on work that has to run again tomorrow, and the day after. We build the loop so code does that job, not you. The work runs on its own, every time, instead of only while you're sitting there babysitting it.

A chatbot answers when you ask. A business process has to run on its own, handle the ugly cases, and not break when something changes. That takes structure: code that routes each step, checks the work, and recovers from failure. AI is one part inside that, pointed at the narrow jobs it's actually good at. Running your operations through a chat window isn't running them.

Zapier, Make, and n8n are boxes of triggers you wire together in yet another dashboard. Another login, another subscription, another screen to babysit. That's the old way of working, and we don't add to the pile. We build the system itself: a leader written in code that routes every step, holds state from start to finish, recovers when things break, and validates every input and output. No dashboard to log into, no one sitting there driving it. Just systems working in harmony, getting the work done.

No. AI on its own is hard to control, so we don't hand it the wheel. Deterministic code runs the loop, and we aim AI only at what code genuinely can't do, under strict guardrails. Reliable first, clever second.

Recurring work that follows a repeatable shape and eats real hours. Prospecting, contact enrichment, content publishing, social posts, back-office admin. If a person does it over and over, it probably belongs in a loop.

No. We scope the tools you already work in and build around them where it makes sense. The difference is the manual part: where you used to spend hours in a tool to get a task done, it now just happens, without you driving it by hand. And if a better setup would genuinely serve you, we'll tell you and recommend it.

It depends on the job. A single task workflow can ship in a few days and stays under €5,000, small businesses included. We've built these before, so we move fast. A full department like sales or marketing, or the central nervous system of a company, is a bigger job: many tasks that have to talk to each other and orchestrate. A build like that takes three to six weeks, and while the price depends on scope, no one's selling a kidney for it.

Every input, output, error, and edge case is checked before the loop moves on. The leader recovers from failures and holds state. We test on synthetic data and ugly edge cases until it holds, then we watch it run and sharpen it over time.

Your call. We can hand over a finished system, live and running, and you take it from there: watch it, maintain it, improve it. Or we stick around. We believe in an always-improving process, so we like staying close to make sure that keeps happening. That part is optional. The retainer depends on the case, but for a single workflow it's usually no more than a few hundred euros a month, and it scales with how much the work actually needs, on-call included.