Glossary

Every word on this site, and what it returns.

A word that needs explaining is a word that costs a sale. For each one: what it is, why it weighs on your accounts, and what it changes at your company.

What we build

AI agent

A piece of software that does a task in place of your teams, from start to finish: read the documents in a file, gather the data, draft a first version, chase a customer. It works inside your own tools, and a person signs off before anything goes out.

Why it matters

A tool helps your teams work faster. An AI agent does the work. The difference shows up on the payroll line, not the software licence line.

At your company

Take the task three people do every Monday morning. The AI agent does it overnight; on Monday your three people review and decide. You do not gain ten per cent on their week, you get their Monday morning back.

AI agent foundation

The plumbing shared by all your AI agents: the access to your tools, the security rules, the record of what has been done. It is built once.

Why it matters

Without it, every AI agent is rebuilt from scratch and you pay three times for the same thing. With it, the cost falls with each new agent.

At your company

The first AI agent carries the foundation and costs full price. The second plugs into it. That is what turns a one-off project into a permanent capability, and it is why we never start with the biggest case.

Guardrail

A limit written into the AI agent: what it is not allowed to do, the amount above which it stops, the case it must hand back to a person.

Why it matters

It is what makes an AI agent deployable. Without guardrails your legal team blocks it, and they are right to.

At your company

The AI agent chases your customers, but never above a given amount, never on a disputed account, never twice in the same week. Those rules are written down and shown to your teams before go-live: that is what makes them use it instead of working around it.

How we work

Diagnostic

Ten days during which we look at how the work actually gets done at your company, and we work out what it costs in hours and in euros.

Why it matters

It is the only point at which you can still say no without having spent anything. Most AI projects fail because nobody costed them before building.

At your company

You leave with a list of cases ranked by what they return, and the discarded cases with the reason written down. If nothing pays for itself, we say so and you stop there. It can be bought on its own.

Framing

The point at which we write down what will be done, by whom, by when and for how much, before anyone writes a line of code.

Why it matters

It is your protection against drift. Whatever is not in that document is neither in the project nor on the invoice.

At your company

Your teams know what they are getting and when. Your finance director has a firm figure. And when somebody asks « could we also… », the answer is a new costed line, not a delay.

Scope

The line between what the AI agent does and what it does not do, settled at framing.

Why it matters

A narrow scope that holds beats a wide one that is vague. Vagueness is the first cause of a project that never lands.

At your company

Instead of « the AI agent handles invoices », we write « the agent processes supplier invoices in PDF from your first twelve suppliers ». The next twelve are a separate, separately costed step. It is less impressive in a meeting, and it is what works.

Going live

The day the AI agent works on your real files, not on a sample, and your teams rely on it for good.

Why it matters

It is the only date that counts. Until it has passed, your investment has returned nothing.

At your company

It is written into the framing, with the real volumes the AI agent will have to carry that day. An agent that copes with fifty test files and falls over on five hundred is not taken live, it is pushed back.

Adoption

Your teams using it every day with nobody standing over them. It is not training: it is the result training is after.

Why it matters

This is where almost all the value is lost. An AI agent nobody uses costs exactly what it cost, and returns nothing.

At your company

Ten days per team: we work alongside the people who do the task, we fix what gets in their way, and we leave when they use it without us. The sign it worked is that they come asking you for the second AI agent.

Pilot

A limited trial, run alongside the real work, to see whether it works.

Why it matters

Most stop there: they prove something and change nothing. A successful pilot that never goes live is an expense, not an investment.

At your company

If you have run one, you know the script: a fine demonstration, an enthusiastic steering committee, and six months later nothing has moved. We do not sell pilots: the first case goes live, or it does not get built.

What we measure

Billability

The share of your teams’ time that ends up on a customer invoice. The rest you pay for yourself: internal meetings, retyping, formatting.

Why it matters

In any business that sells time, this is the figure that governs the margin. It moves slowly, and every point recovered drops straight to the bottom line.

At your company

A consultant bills on average 66% of their time where the sector aims at 75%. Across twenty-five people, those nine points are 3,600 hours a year, or 360,000 euros at a hundred an hour. Recovering a third of it is already a good result.

Point of billability

One per cent more time billed.

Why it matters

It is the unit we think in, because it converts into euros with no argument possible.

At your company

Across twenty-five people at sixteen hundred hours a year, one point is four hundred hours. Redo the sum with your headcount and your hourly rate: you get the figure an AI agent has to return to deserve being built.

Rework rate

The share of the AI agent’s output a person has to do again behind it.

Why it matters

It is the figure that says whether the AI agent really saves time. At fifty per cent rework, it saves none.

At your company

We measure it before and after, on the same files. That is what lets you say « the AI agent saved us this many hours » without anyone round the table having to take your word for it.

Quotation

The price, worked out after framing and not before, from what has been decided.

Why it matters

A price list would have you pay the average of everyone else. You pay for your case.

At your company

An AI agent that reads three invoice formats and one that reads thirty do not cost the same. The diagnostic tells you which one you are before you commit to anything.

Finding customers

Outbound

Going after customers one by one, by message or by phone, instead of waiting for them to come.

Why it matters

It is the fastest channel to start and the most expensive in human time. Which makes it the first place an AI agent pays for itself.

At your company

If your prospecting rests on your partners’ time, the most expensive time in the firm goes into looking up contacts and drafting messages. The AI agent does the research and the first draft; your partners keep the conversation.

SEO

The work that moves your pages up the Google results.

Why it matters

Slow to take effect, lasting once it does: it is the channel that keeps working when you stop paying.

At your company

It needs steady output, and that is exactly what gives way when the team is busy elsewhere. The AI agent produces the first drafts; you keep the review and the sign-off.

AI answers

What ChatGPT and its like reply when someone asks them for a supplier like you.

Why it matters

Your customers already search there. You do not show up by the methods that work on Google, and almost nobody has worked on it yet.

At your company

This is the channel where the lead is taken now, and it is measured in citations, not positions. We look at what is said about you today, and at what has to be published for that to change.

CRM

The central file holding your customers, your prospects and the history of your exchanges. Salesforce, HubSpot and Pipedrive are examples.

Why it matters

It is where manual data entry costs the most, because it is done by salespeople paid to sell.

At your company

Count the minutes each salesperson spends filling in fields after a meeting, multiply by the number of meetings in a week. That is selling time you are paying for as admin.

Contact

A problem first. The tool after.

Describe a situation that costs you time. We will tell you honestly whether AI is the right answer.

Tell us about your project