The AI agency that opens one system at a time.
Every business runs on ten systems — acquisition, sales, delivery, customer relations, finance, legal, HR, IT, management, and the knowledge underneath them all. Most are closed : the work gets done, but by hand, twice, late. We survey the ten, put a figure on what each one costs you, and open only one : the one with the shortest payback. The survey is free and delivered within 72 hours.
The tasks we remove first
What your ten systems cost, and what they give back.
Survey of a manufacturing SME with 84 staff, before and after opening four systems out of ten. Your figures will differ : that is precisely what the survey is for.
Four ways to open a system.
The same method every time — survey, one system, measure — and four kinds of work depending on what the survey finds. None of them starts with a tool.
The survey of the ten systems
One workshop on your premises, ten systems rated closed, ajar or open, and a figure on what each one costs you. Free, delivered within 72 hours, and yours to keep even if we never work together.
Agents that read your documents
An agent that answers from your quotes, your procedures and your exchanges — and cites where its answer comes from, so it can be checked. It prepares; a person validates.
Automation circuits
Built in n8n or Make, visible and editable by your team. Read an email, extract, search, fill in, notify — and stop when a human has to decide.
The document base
The first job on almost every engagement : making readable what already exists. An agent is only worth what it can read.
What an open system does, step by step.
Nothing magical : a sequence of steps you can read, stop and correct. Here is the most frequent one, quoting, in the order it runs.
Schematic drawing. The real scenario is built in your automation tool and stays editable by you.
Survey, one system, measure. In that order.
We never start with a tool, and we never open ten systems at once. Every engagement follows four steps, and the third one is the one that most agencies skip.
You describe the task that costs you the most. We tell you whether it is for us. If not, you leave with the two reasons why.
Ten systems rated, a figure on each. Yours to keep, whatever happens next.
The one with the shortest payback. Document base first, then the agent or the circuit, then a person who validates.
Announced at the start, measured at the end. If it has not moved, we say so.
The tasks that cost, industry by industry.
The areas where we are called most often.
What we are asked before we start.
Do our documents leave the company ?
Only if you decide so, and it is the first decision of the workshop. Depending on sensitivity, we install on your premises, in an isolated European hosting, or we use an online model. The choice is written into the quote. Nothing is used to train a model.
We tried ChatGPT. It made things up.
That is the normal behaviour of a model given nothing to read. A useful agent works on your documents and cites its sources : you check in one click instead of taking it on trust.
How long before it is useful ?
The survey is delivered within 72 hours. The first system opens in two to six weeks, depending on the state of your documents — that state, not the technology, is what makes the difference between a short project and a long one.
What do you not do ?
Three things, written on every quote : no automated rejection of a person, no legal or financial advice in place of your adviser, no content published without a named human review. They are not disclaimers. They are the three places where projects go wrong.
You are in France. Do you work in the UK ?
Yes. The survey workshop is on your premises; everything after is run remotely with a weekly call. We work in English, and the documents stay where you decide.
We will tell you which system to open first.
Describe the task that costs you the most. We will tell you what is feasible, how long it takes and what it costs. If the answer is no, you will leave with the two reasons why.