We automatewhat’sexpensive.And we build what isn’t there yet.

AI when the problem calls for it, automation when automation is enough. The point is never the technology. It’s the result.

Request a diagnosticRequest a callback
Scroll
Project
Your business
Drawn by
C. Torr
Units
px
Place
Toulouse. Remote anywhere.
Local time
--:--
Sheet
01 / 08
What we do

We start by
sorting it out.

Not everything is worth the same. Some of what you run should be replaced with something robust, some of it never existed and has to be built, and some of it is not worth touching. That sorting happens on your site, inside your processes, with your teams. And we tell you where AI actually wins, and where a plain automation is enough.

What we replace

Things that already work, badly, and that everyone has learned to work around.
  • The spreadsheet everything depends on, that nobody dares touch.
  • Re-keying from one tool into another, every morning.
  • The approval sitting three days in an inbox.
  • The month-end check that finds the errors too late.
  • The whole process that lives in one person’s head.
What you get back is the same thing, made robust. Monitored, documented, and it does not fall over on a Friday night.

What we build

Software written for you. Which is to say: the thing you need and cannot buy anywhere.
  • An application cut to your trade: CRM, client portal, back-office, case tracker.
  • An engine that prices, compares or calculates what only you know how to work out.
  • Chains that connect your tools to each other and run without anyone.
  • Understanding what arrives as prose: email, contracts, files, requests.
  • Listening to what gets said in your business, and turning it into numbers and actions.
  • Agents that answer, qualify and follow up. In production, and watched like everything else.
  • Training your teams on your real accounts, not on textbook cases.
At that point it stops being about saving time. It is about doing what you could not do.
Timelines

Months?
Not any more.

With the right tools and know-how, quality, production-grade software gets built in days. The time goes somewhere else.

We take the time to understand what you need, on the ground, with your teams.
We work the specification until it leaves nothing to chance.

Implementation isn’t the hard part any more. Knowing exactly what to build is.

The method

Three stages.

01 / THE DIAGNOSTIC

We come and see how you actually work.

One to three days, on site. We watch what people really do, not the org chart. You leave with a written plan: what, in what order, what it costs, what it returns, and what isn’t worth touching.

The plan is yours. Even if you never call us back.
02 / THE BUILD

We build it, and we put it in production.

Days or weeks, not months. At the end it runs for real, not as a demo. Your teams are trained on it, everything is documented, everything is visible.

Scope and price fixed before we start.
03 / AFTER THAT

We watch it. You stop thinking about it.

If something breaks, we are the ones who tell you. Not your customer. And you can stop whenever you want: nothing collapses when we leave.

No minimum term.
One example, in detail

What they could not
do. And now do
every single day.

A business where the sale happens on the phone. The calls had been recorded for years and nobody listened to them: an attentive manager hears five in a good month. That was not wasted time. It was a capability they did not have.

400,000+calls heard and analysed
11,000+actions taken out of those calls
1,800+training runs played by the teams
16,000+SMS replies sent with nobody there
A rep can replay a call they lost.From the exact minute it turned, against a customer who remembers what was just said. This is not a faster version of what they did before: before, it did not exist.

It all runs on their accounts, in their database, under their name. Every analysis can be replayed, every version of every instruction is dated, and when a stage is skipped the system says so instead of pretending.

The guarantee

You
keep
the keys.

Here is the question nobody asks in the first meeting: what happens the day we stop working together? This is our answer, in writing, before anything starts.

The code and the configuration are yours, in your repository.
Everything runs on your accounts and your credentials. Nothing in our name.
You can see what ran, what failed, and what it costs.
A written exit plan: how to take it back over, who to call, what to do.
The documentation is written for your team. Not for us.

This is not generosity. It is the condition for owning the thing rather than renting it.

Who we are

Callum
Torr.

Forward deployed engineering: automation and applied AI, built inside your systems and left running. Based in Toulouse, working on site and remote.

For ten years we ran engineering teams. Which means: we were the ones woken at three in the morning when the system went down. You learn quickly, in that job, to tell the difference between what works in a demo and what is still standing two years later.

The last five we spent building AI systems in production: agents, call analysis, AI grounded in company documents. Enough to know when AI settles a problem and when it adds one.

Today we don’t run teams. We build, we ship, and we stay while it runs.

Forward Deployed Engineer is the accurate title for it: an engineer who works inside your processes, not alongside them. In your repo, in your stack, on the hook for what happens after go-live.

Alongside client work, we build our own tools for agents: their memory, a map of what a company runs, contracts between agents that don’t know each other. Three are already public, and we run all of them ourselves before we put them in front of you. See our tools at cloud.montytorr.com →

Contact

Let’s talk.

Tell us in two lines what is eating your time. We reply within one business day.

Your details are used only to reply to you. They are not sold and not used for prospecting.