What we won’t build

Most studios list what they do. The refusals tell you more. These come from years spent on the other side of the counter, in complaints and vulnerable-customer teams at a regulated bank.

Anything that hides a human from someone who needs one

A chatbot that can’t hand over, or is built to stall, is a wall dressed as a door. Every assistant we build has a clear route to a person, and it takes that route on its own when someone is distressed, confused or asking for it.

Retention flows designed to trap

Cancel buttons that don’t cancel. Offers that reset the clock. Six screens to leave and one to join. If a customer wants to go, the system lets them go. Your renewals should come from the product being good.

Automation that removes the check without removing the risk

Speeding up a process is only a win if the decision inside it was safe to automate. Refunds, credit decisions, anything touching a vulnerable customer: we build the fast path and the stop, and we tell you which is which.

Anything that pretends to be a person

Our assistants say what they are. No fake names, no fake typing delays to feel human, no “Sarah from support”.

Scraping, and data you don’t have the right to use

If the lawful basis isn’t clear, we won’t build on it. We’ll help you find one, or find a different approach.

Work that isn’t worth doing

The most common outcome of a scoping call is “don’t automate this yet.” We say so. A plan that starts with the wrong system is money you’ll want back.

If something you need isn’t here and isn’t on the services page either, ask. Off-list is fine. Off-principle isn’t.