First call
I get to know your business and spot where you lose time and money.
- No commitment, no card
- I also tell you when it is not worth it
- You leave with concrete leads
I start by understanding how you actually work. Only then do I build the automation that saves you time and money — the one that fits your business, not a catalogue.
Fifteen years of infrastructure, put to simple use
n8n for workflows, large AI models for understanding, Docker and Linux to hold it all together
Of my own workload saved through AI
Years running production systems
Deployments delivered
Infrastructure projects
Uptime on the services I operate
The shapes differ from one company to the next, but the causes always come back. Here are six — you will likely recognise several.
Anywhere the same information is typed twice: from the form to the spreadsheet, from the spreadsheet to accounting, from accounting to email. That is pure lost time, and where errors are born.
{
"node": "Edit Fields",
"mode": "manual",
"assignments": [
{ "name": "client", "value": "={{ $json.societe }}" },
{ "name": "montant", "value": "={{ $json.total_ht }}" }
]
}What should go out on a fixed date and slips through because nobody thought of it. Follow-ups, reminders, renewals, periodic checks: a machine never forgets.
{
"node": "Schedule Trigger",
"rule": {
"interval": [
{ "field": "weeks", "triggerAtDay": [1], "triggerAtHour": 8 }
]
}
}Files that sit idle because they are waiting for one specific person to take one specific action. Often, alerting the right person at the right moment frees up entire days.
{
"node": "Wait",
"resume": "webhook",
"options": { "limitWaitTime": true, "resumeAmount": 48 },
"onTimeout": "Relancer le responsable"
}Time spent finding a document, checking a figure, cross-referencing two sources. When data is filed and reconciled automatically, that search disappears.
{
"node": "AI Agent",
"promptType": "define",
"text": "=Extrais le client, le montant et l'echeance de : {{ $json.texte }}",
"hasOutputParser": true
}Everything that always follows the same rule: sorting, routing, calculating, approving below a threshold. Your rules, applied tirelessly — and you take back control the moment something falls outside them.
{
"node": "Switch",
"rules": [
{ "outputKey": "auto", "condition": "montant < 500" },
{ "outputKey": "validation", "condition": "montant >= 500" }
]
}An automation that dies quietly costs more than no automation at all. Fifteen years of infrastructure serve one purpose here: making sure what I deliver lasts.
{
"node": "Error Trigger",
"onError": "continueErrorOutput",
"then": ["Alerte Slack", "Journal Postgres", "Reprise automatique"]
}A real workflow: information comes in, AI analyses it, your rules apply, and you approve where it matters.
Every workflow is different: this one is an example. Yours will be built from your own tools and rules.
Four steps, with no commitment to begin with.
Thirty minutes together, free. You walk me through your days, your irritations, what breaks when someone is away. I spot what can disappear — and I also tell you what is not worth it.
For each task: what it really costs you, what automating it would cost, and how fast it pays back. We start with the one that pays back soonest, not the most impressive.
First version in one to two weeks, built on your rules and plugged into your existing tools. You use it for real, we adjust what does not fit. Nothing goes live until it suits you.
Monitoring, backups and fixes. You are warned before something breaks, not after. And we move to the next task when you decide.
Nothing is fixed in advance: the price depends on what we find together. But you know from the start how we proceed, and you never pay before you have seen something.
I get to know your business and spot where you lose time and money.
A real workflow designed for you, a live demo, and the return-on-investment figures.
Building, deploying and documenting your automations.
So your automations keep running, and evolve with you.
Indicative prices excluding tax. The exact amount is given in writing after the full audit, before any commitment. The audit fee is deducted if you go ahead with the build.
My own lab runs on the automations I design. That is where I put them to the test before offering them to you — and the best guarantee that I know what I am talking about.
AI agents that know my entire environment, the software inventory and its versions, with persistent memory. Routine operations take a fraction of the time they used to.
Workflows take care of routine upkeep on my infrastructure. What had to be checked, sorted or handled by hand now happens on its own, and I only step in for exceptions.
Several automations run permanently in my own lab without my thinking about them. They are the same mechanisms I put in place for clients, adapted to each context.
I do not know your business better than you do. My job is to look at it closely, ask the questions you stopped asking because you have always done it that way, and spot what can disappear. Only then does the technology come in — and it adapts to you, never the other way round.
Two companies in the same sector do not work alike. I do not copy a solution seen elsewhere: I build yours, from your rules and your tools.
A human approval step stays wherever the decision commits you: a quote, a client reply, a payment. Automation prepares; it does not decide for you.
Everything I build belongs to you, documented and transferable. If you leave, it keeps running without me.
If your question is not here, just ask: the answer is free.
Tell me in a few lines what is costing you time. Reply within one working day, or book a slot directly.