Install AI in habits, not in the tech stack.
A targeted engagement to turn an intention, "we should do something with AI", into real, observable use that survives my departure.
Who this is for
You run a 50 to 300 person company. You've tried. It hasn't taken.
The typical scenario: a curious manager pushed for the company to get started. You bought a few ChatGPT or Copilot licences. A training session was organised. Three people still use it. The others closed the tab.
You know AI is going to matter. You don't know whether you're looking at the right tool, the right use case, the right moment. And you have neither the time nor the appetite to make AI a project of its own that distracts the organisation from what it actually needs to do.
This is exactly when I step in.
The engagement
Adopt · Adapt · Automate
Three phases, in this order. Not the reverse. The most common trap is to start with automation because it sells best in demos, while no one in the company yet uses AI day to day.
Phase 01
Adopt
Individual reflexes
Phase 02
Adapt
Individual reflexes
Team workflows
Phase 03
Automate
Individual reflexes
Team workflows
Autonomous systems
AI in your teams' hands
Diagnose the tasks where AI delivers an immediate gain, the ones your teams perform every day, not textbook cases. Install the tools. Train directly on real files, real emails, real analyses. The objective isn't that they know how to use AI. It's that they use it spontaneously.
AI configured for your trades
Once usage is settled, we customise. Specialised assistants for the key functions. Libraries of prompts validated by the trades themselves. Workflows that fit into existing tools: Outlook, Teams, your ERP, your CRM. AI stops being a separate application and becomes a layer in daily work.
Automate, when it's ready
Once habits are anchored and we know precisely where AI creates value, we automate. Agents for repetitive tasks, integrations between systems, supervised autonomous flows. Not before. Automating a process no one yet masters manually is building a motorway without knowing the destination.
Success is measured by observed adoption, not declared.
What you keep
By the end of the engagement, here's what stays.
Not a report. Not a slide deck. Living assets that belong to the company and continue to work without me.
01
A map of real-world use
Where AI delivers gain, where it doesn't, in your specific operations. Not a theoretical matrix, a document grounded in what actually worked at your site during the engagement.
02
Configured assistants and prompts
For each key function: a calibrated assistant, a library of tested prompts, integrated workflows. Everything documented so your teams can keep evolving them on their own.
03
A trained team, not an instructed one
The distinction matters. Your people won't have learned what AI is. They'll have learned to use it on their own files, with the reflex to come back to it.
04
A synthesis document
The story of the engagement: what was installed, what worked, what resisted, and why. Short, dense, for your internal use. This is what makes me redundant. The objective from day one.
How it works
The engagement, in practice.
An AI Adoption mission is not an audit, nor an IT project, nor a training programme. It's a short, intense intervention done with your teams, not for them.
Cadence
1 to 2 days per week
On-site and remote, alternating. The rhythm adapts to the phases. Denser at the start, more diffuse toward the end of the mission.
With whom
A sponsor and a working group
You, or a member of the executive committee, as sponsor. A working group of three to eight people from operations who do the real work.
What I need
Access to operations
Not staged interviews. Time with the teams, on real work, with the freedom to test, break, adjust.
Pricing
By engagement, not by the day
A fixed price agreed after scoping, based on the perimeter. No hourly billing, no surprise change requests.
Filter
What this engagement is. What it isn't.
It is
- A short, focused intervention that ends.
- Work with your teams on their real files.
- An engagement designed to make me redundant.
- A deliverable measured by usage, not by output volume.
- One contact only: me.
It is not
- An 80-page AI maturity audit.
- A comparison of forty tools you'll never use.
- A generic training programme.
- A keynote for the executive committee.
- A team of junior consultants who move in and stay.
A note on discretion
I don't publish client names. The Belgian SME world is small. When an executive entrusts me with their organisation for eight weeks, they also entrust me with what isn't working, what they wouldn't say in a board meeting, and the human fragilities that make this work interesting.
That discretion is part of what you're buying. If you need logos to validate a choice of consultant, I'm probably not the right profile, and that's perfectly fine.
What's next
If this resonates, write to me.
Two sentences are enough: where you are, what you'd like to unlock. I respond personally, within 48 hours.
Write to me