Opinion · Agentic AI & Lean Manufacturing · May 2026
I’m just back from Hannover Messe 2026 with a conviction that grew stronger at every conference: the race to robotize is not the only path — nor the most relevant — to transform our companies. I believe far more in agentic AI. And it’s precisely because I come from Lean that I say so.
Here is why — and how, concretely, this technology is already changing the game, both for me and for my clients.
❄️ A sentence that left me cold
At Hannover, I heard this sentence, spoken as if it were an obvious truth:
« Labor costs are so expensive. »
And around the table, at the end of the conference, everyone nodded along: robots are the solution. In their companies, everything is going well. Really?
Only one speaker dared to challenge the consensus: no, robots will not be so easily accepted, and the cultural change will be difficult — very difficult, depending on the country.
I found him courageous. Because what I heard that day was a truncated reading of performance: we look at only one side of the equation — cost — without ever putting on the table the value that humans create.
💰 An employee’s cost is only half the debate
When a client tells me “it’s too expensive”, my answer is always the same: and the value it brings you — when do we talk about that?
It’s exactly the same mechanism: yes, an employee has a cost. But what is the value they create? What is their ability to decide, arbitrate, solve the unexpected, and lead a team?
A Lean saying keeps coming back to me: “Either you invest now — or you pay ten times more later, the price of the losses you could have avoided.”
Believing that you’re investing by replacing a human with a machine is solving the wrong equation. The losses you don’t see right away always end up being paid back a hundredfold.
🧠 Lean has always put humans at the center
I spent years driving operational excellence, then teaching it as a consultant. And if there is one thing I have learned:
The entire Lean toolbox has a single objective → to support humans so that they use their intelligence to make decisions, at every level of the company.
Not to constrain them, even less to replace them. But to free them from repetitive tasks, and give them back the time to devote their intelligence to what truly creates value.
This is exactly what agentic AI enables. And this is why I believe in it.
🛠️ Example #1 — My “mileage expenses” agent, built yesterday
I’m an independent consultant. Every month, two tasks devour my time: mileage expenses and accounting.
Yesterday, I built an AI agent that:
- 📍 Reads my GPS coordinates as soon as I get in my car
- 🏷️ Asks me at the end of the day: business trip or personal?
- 🧮 Calculates the kilometers driven every month
- 💶 Applies the official tax-deduction rate
- 📄 Writes the invoice in my accounting software, with the detailed Excel as an attachment
Build time: a few hours. And the longest part wasn’t even building the agent — it was upgrading my IT setup to use the best tools available.
My first impression? “Whaooooo.”
The next agent? My entire accounting workflow. 🚀
🏗️ Example #2 — In industry: managing prevention plans
Do you manage a lot of subcontractors? Then you definitely have a prevention plan challenge: copy-pasting from obsolete versions, sites starting without a signed plan, operational tasks described in the plan that no longer match at all what the subcontractors actually do on site, and managers complaining they no longer have time to run proper site visits…
Imagine an AI agent that:
- Creates a blank plan as soon as the offer is validated
- Pre-fills the blocks from the offer’s content and from the latest central-services templates
- Notifies the manager + their N+1, with the incompressible signature lead times
- Launches the signature workflow once filled
- Compares to previous versions and computes the % of modifications
- Triggers an alert if the weight of modifications raises concerns
Result:
- ✅ No more copy-pasting from obsolete versions
- ✅ No more sites starting without a prevention plan
- ✅ No more gaps between the signed plan and what is actually being done
- ✅ Targeted audits on sites most likely to be at risk or delayed — no more blind controls
- ✅ A real management tool: you finally know where to focus your coaching effort
- ✅ All these indicators flow into Power BI — aggregated by service, department, plant, country, continent. This bird’s-eye view is the very condition of steering. Because making things visible is the prerequisite to any performance.
And above all: time freed up for what really matters.
Today, some managers ask to go back to operational roles — too many office tasks, not enough field time. With agentic AI, we put managers back where they belong: not behind a screen, but on the ground, alongside their teams.
🎯 And what about simulation in all this?
My simulation tool PLANORA already turns Excel data into optimized production schedules. The bottleneck today? Preparing the Excel files in the right format.
With agentic AI, this bottleneck disappears: the agent pulls the data from the ERP, formats it, feeds PLANORA, and delivers the optimized schedule. A true “one-click planning”. Fully autonomous.
Agentic AI does not replace simulation. It finally makes it accessible day to day.
🇫🇷 Why I believe in it more than in robots
In a country where cultural transformation is complex, trying to robotize everything means exposing yourself to losses you will pay back a hundredfold later.
Agentic AI, on the other hand, doesn’t ask operators to coexist with a machine. It silently offloads time-consuming, low-value tasks. It frees up time for what makes the value of a French company: know-how, judgment, collective intelligence.
By design, it is aligned with the Lean philosophy: place humans at the center, free them from the superfluous, give them back their decision-making power.
📩 And you?
- What are the 3 processes that take up the most of your time in your daily work?
- Have you ever identified, within your teams, repetitive tasks that drain energy without bringing value?
- What cultural barriers do you encounter when you talk about AI with your colleagues?
Write to me. I would be delighted to discuss it — or even to build a first agent with you in a few hours, as I did for myself.
Because the real revolution isn’t in the robots we see. It’s in the agents we don’t see — and that bring humans back to their craft.
📩 Let’s discuss your project
A question, a need, an idea for an AI agent to build?
Write to me — I would be delighted to discuss it.
