Where most stop
Marginal productivity
Copilot, ChatGPT in teams. Useful, but changes nothing in the structure of the company.
02 · My method
Not all "AI" is created equal. A grid to tell them apart.
I read AI initiatives on two axes. What they touch: a tool, or the company's business model. Their horizon: tactical gain, or strategic transformation.
The two-axis grid
Tactical
Strategic
Business model
Business model · Tactical
Hybrid zone
New AI-based service, limited scope
Touches the business model at small scale. Interesting if it's strategic learning in disguise.
Business model · Strategic
Where I work
Business model transformation
What changes your unit costs, your offering, your competitive advantage. The real topic, rarely addressed.
Tool
Tool · Tactical
Tool · Strategic
Hybrid zone
Tool-as-lever
A well-chosen tool on a critical process can become strategic. The exception, not the rule.
My work is to orient your attention toward the top-right quadrant. Productivity counts and I don't ignore it; I just make sure it doesn't obscure the central question: what does AI change about your business model?
Three questions, in this order
1
Axis 1 · The business model
Where is value created?
Where is it lost? What does AI make technically possible that wasn't possible yesterday?
OutputA strategic thesis.
2
Axis 2 · The data
What do you already know?
About your customers, your operations, that you don't exploit? What data is missing to decide better?
OutputAn asset map.
3
Axis 3 · The organization
Who is ready to move?
Who will resist? What level of autonomy can your team carry on this topic?
OutputA deployment plan.
The three axes aren't handled separately. They confront each other. An ambitious strategic thesis without data assets is wishful thinking. Data assets without an organization ready to move is a dormant deposit. It's the intersection of the three that surfaces the initiatives that matter.
From the grid to the field
When you arrive with a list of tools, or with no idea where to start.
Most requests I receive don't talk about business models. They talk about tools, or about a blur: automate quotes, follow-ups, or simply "we need to do something with AI". I run all of them through the same grid: what protects or moves revenue, what touches the cost of service, what only saves hours.
That is exactly what the ordre de marche does: three weeks to sort a list with this grid, and come out with three workstreams.