Autonomous Intelligence
Talk to us

Made-to-measure industrial AI

AI that works for your teams, inside your plant.

The AI prepares the work, your people decide. By default, it runs on your own servers.

Talk to us What we do

Most plants run in several eras at once. Yours probably does.

Illustration: an engineering office buried in drawings and requests for quotes
Illustration

2.0 Manual

The plan lives on a whiteboard.

Plan
On a whiteboard
Machines
Set by hand
Data
On paper
Decisions
By experience
Illustration: a team reviewing a document-to-ERP workflow on screens
Illustration

3.0 Automated

The machines run programs. The office runs the ERP.

Plan
In the ERP
Machines
Run programs
Data
In separate systems
Decisions
From reports
Illustration: a digital view of a shop floor with machine data
Illustration

4.0 Connected

Plenty of data. Little action.

Plan
In the ERP
Machines
Connected
Data
Everywhere, on dashboards
Decisions
Still by hand, late
Illustration: engineers beside a server rack on the shop floor
Illustration

5.0 Assisted

The AI proposes. Your people decide.

Plan
Proposed by the AI
Machines
In step
Data
Used as it arrives
Decisions
Yours, with a proposal ready

Start with one painful workflow. Proof on your own files in about two weeks. Then scale.

  • About 20%

    less energy, by our estimate

    Manufacturing plant, Europe

  • One truck fewer

    than the logistics team’s own plan, on one real delivery day replayed with their constraints

    Distributor, the Americas

  • 30%

    fewer early tool changes, with vibration sensors we installed

    Machining, Europe

Projects in 12 countries. No client names, no logos: we protect the companies we work for.

What we do

A hive of specialists

We started small, and a bee is small and useful.

Founded in Europe in 2019. In Brazil since 2022. Present in France, Romania, Singapore and Brazil.

  • Laurent Poyet CEO Formerly led international programmes in automotive safety.

Our 37 specialists, from the sensor to the decision.

Data
Scientists
Engineers
Analysts
Plant
Process engineers
Quality engineers
Operations specialists
Systems
ERP specialists
Architecture engineers
  • Integrated AI platform
  • Cost reduction
  • Quality
  • Less downtime
  • Flows and logistics
  • Energy and ESG
  • Compliance and sovereignty

Where AI pays back first

Six places where we suggest starting. Each one is built on your files, your systems and your rules. Real payback depends on your volumes, so we measure it before any pilot.

Expected payback in weeks

  • Orders and quotes, without re-typing

    An agent reads each email, PDF or WhatsApp and prepares the order or quote in your ERP, for your team to review and send.

    Built for one client: a WhatsApp assistant that identifies each contact and passes the conversation to a salesperson.

  • Tenders checked against what you make

    The AI marks every requirement as met or missing and lists the certificates to attach, so you can decide whether to bid.

    Start with 15 to 20 recent tenders, won and lost.

  • Paper and photos into your system

    A phone photo becomes the fields your system needs, and doubtful cases go to a person.

    In production, on a client’s own server.

Expected payback in months

  • Truck loading and routes

    Plans built with your real constraints: truck types, time windows, loads, every stop.

    In production. On one real delivery day, replayed: one truck fewer than the team’s own plan.

  • Maintenance at the right moment

    Vibration sensors we install, and a model that learns each machine.

    30% fewer early tool changes.

  • Energy per line and per product

    Our own meters on each line, and a production plan that follows what they measure.

    About 20% less energy at one plant, by our estimate.

By default, your data stays inside your walls.

Our models run on your servers by default, and your know-how stays yours.

Stays inside

  • Production data and sensor signals
  • Drawings, quotes and prices
  • Documents and photos
  • The models we train for you, and what they learn

Leaves only if you decide

Data sent to an outside service, such as a hosted language model or a traffic API, if you choose one. One assistant we built for a client uses a hosted model, because the client chose it.

We build on what you have

On the line, at the edge and through APIs to your systems: SAP, TOTVS, your MES, your data platform. If you already run your own AI, we build on it.

We start small, on your own files.

No audit. No report. A result.

  1. One call

    A conversation

    You tell us what costs you most. If everything runs well, we don’t invent a project.

  2. About 2 weeks

    Proof on your files

    15 to 20 of your own documents, under NDA. You see what the AI gets right and what it hands to a person.

  3. 4 to 6 weeks

    A pilot

    Fixed scope, fixed price, on one workflow, on your systems.

  4. Then scale

    Production, then the next workflow

    By default, it runs inside your plant. We measure the payback.

Back to What we do

Projects

A selection, by problem. Each was built for one company; yours would be built for you. No client names: we protect the companies we work for.

    Let’s talk about your plant.

    Tell us the workflow that costs you most, and we’ll tell you honestly whether AI pays back there.

    We use these details only to reply and follow up. Privacy notice