The work you would not dare give to a machine: AI where it pays, your expertise where it counts.

A model that handles one file well does not handle a hundred in a row. On business tasks, success falls from about 58% in a single turn to 35% across several. What gets through is an answer given to a client with complete assurance, and wrong.

AI drafts, a person re-reads: across 106 experiments, the pair does worse on average than the better of the two alone, because the human follows the machine when it is wrong, and training does not change that. The authors point to another route: give the machine only the parts where it is clearly better. Three experiments out of 106 had tried it. The other option, nobody re-reads and nothing is counted, has not even been measured: you do not know what goes out.

Where a file passes through someone before it leaves, it spends most of its life waiting. Across 2,501 medical exams, turnaround fell from 60 to 48 minutes while working time did not move: only the wait went down. On business documents, Microsoft measured a 6 to 12% gain, and nothing on those that take more than a week.

To move a lead time, you need to know which of the ten files still needs a judgment call. Writing faster changes almost nothing.

Whoever answers before you wins the job, even when they are worse: before signing, the client checks the price and the lead time; the quality, they find out afterwards.

“Why was this job priced that way?” — Ask Monique. “This file, how do we handle it?” — Ask Michel.

There is a Monique and a Michel in almost every company, and rarely in the same department: one on quotes, the other in operations. Neither was ever given the job of knowing: they know because fifteen years of files have crossed their desks.

We tried writing that knowledge down: forty pages nobody opens. Giving it to a model as is does not work either: it applies the rule seven times out of ten, and on the rare cases it answers wrongly with complete assurance.

So the people who know go back over all ten, because nothing says which three are wrong. Their procedure says what to do; it does not say which files need a judgment call, and nobody ever noticed. That is where their week goes, on seven files nobody had asked them to look at. And in the company next door, where everyone builds and signs their own files alone, nobody goes back over anything. All ten go out, the three with them.

When the right way to handle a file is known to only two or three people, the company cannot work without them. It cannot teach it to those who arrive either. Each newcomer relearns it alone, waiting for their opinion, and what they learn leaves with them if they go. That is the only real urgency, and we are not the ones inventing it.

Seven times out of ten — SOP-Bench (2025): over 2,000 tasks drawn from procedures written by specialists, across 12 domains; the best model-agent combination reaches 72.4% success.

Your best people cannot be everywhere. The way they work can.

Once it is in the software, that way of working serves the newest hire as much as the old hand, and from one branch to another the same job gets the same answer. In a customer service team that had trained its software on the work of its best people, beginners handled 34% more files an hour, and the most experienced gained nothing.

A team that aligns on a model from the market ends up answering like its competitors. A team that aligns on its own way of working looks like nobody but itself. Three years from now, your competitors will have the same models as you. Not your way of working.

(Brynjolfsson, Li & Raymond, Quarterly Journal of Economics, 2025 — 5,172 agents, 3 million conversations.)

Part of your know-how is written down. Your company reads those texts in its own way, above all when a case does not fit the box, and that way is written nowhere. That is what we write with you.

1. With the people who know, we go back over jobs already sent out: this one was right, that one we would not send again, and why.
2. The software runs them again and returns two things for each: an answer, and a score for how solid that answer is.
3. We read the table: cut here, and it handles six out of ten, and here is what it misses. You choose where to cut, with the people who know.
4. Below the cut, the software works alone, and a random sample checks its error rate. Above it, the file goes back to whoever can decide it. When the software does not know, it does not guess: it passes the file up.

Your teams stay on all their files. What we take off them is a re-read that was finding nothing.

AI is already in your company, but in the corners: across 11,021 companies, 26% use it and about 1% use it as the tool of a business function (France Num, 2025). We do not take projects whose aim is to cut jobs: they turn on you at the first awkward case.

Sources — CRMArena-Pro, Salesforce AI Research (4,280 tasks). Vaccaro, Almaatouq & Malone, Nature Human Behaviour, 2024 (74 studies, 106 experiments, 370 effect sizes). Batra et al., American Journal of Roentgenology, 2023 (2,501 exams: turnaround 59.9 → 47.6 min, wait 33.4 → 21.4, reading 26.5 → 26.3). O'Neill et al., Radiology: Artificial Intelligence, 2021 (the alert alone changes nothing; reordering the queue: 15.5 → 12.0 min). Dillon et al., Microsoft Research, 2025 (randomised trial, 6,000 employees, 56 companies).

Among French companies of 50 to 249 employees, 31% use AI (INSEE, 11,000 companies). Buying the tool was the simple part. We take on the work itself, around what your teams know.

These measurements come from a radiology department and from a vendor's telemetry, not from a French SME. In your company, we measure before we build.

What our claims rest on →

No — that's the normal situation, and that is exactly the work. If everything were written down and followed to the letter, there would be nothing to take on.

What we need isn't your documentation: it's your past jobs — bids, costings, lessons learned — and half a day a week with the two or three people who know, including the busy weeks. That is, by far, the condition that fails most often.

One practical point is settled in twenty minutes: where those files live. A shared drive, however badly organised, is enough. Scattered across individual mailboxes, no.

A first number from the twenty minutes: what this work costs you today, estimated from your own figures.

At the end of the three reference files, four to six weeks later, you have three numbers measured on your archives and the plan for the project. They are the baseline.

The software's return is measured once it is in place, against those numbers, in the follow-up. The software is built so that it can later extend to a second piece of work.