Everyone talks about AI adoption, but I suspect very few companies are actually learning anything — the companies, not the people.
Yesterday, while discussing AI technologies and their integration into business workflows with Flavio Fabbri and other colleagues, I realized something: even though ChatGPT, Copilot, Claude, Gemini, Cursor, and every other LLM-based tool are now everywhere — and people are genuinely becoming more productive — that improvement stays individual. It never becomes organizational.
Think about it. Using AI doesn’t automatically build shared capabilities or competitive advantages for the company. What actually happens is that someone quietly automates their reports, someone else writes code twice as fast, and a few others use an LLM to sharpen their decision-making. But nobody is necessarily aware of what anyone else is doing — and as a result, the company keeps running exactly as before, just with some (or many) individually “enhanced” contributors.
And sure enough, when organizations speak candidly — away from the spotlight — they admit they can’t measure any real business advantage from AI adoption.
Could it be because this kind of shadow usage lends itself perfectly to duplicated effort and measuring the wrong things entirely?
None of this actually matters:
- how many workflows you’ve built
- how many tokens you’ve burned writing code
- whether chatbot X helped you cut customer care costs by Y%
The questions companies should actually be asking are:
- What are we learning as a company?
- Which capabilities are becoming replicable and transferable across departments?
- Are we needlessly duplicating effort and cost?
- Are we orchestrating an AI strategy in a centralized, deliberate way?
- Are we genuinely training our people to become mini-CEOs of their functions — each managing 30 resources, a mix of human beings and digital agents? (hat tip: Matteo Genova Bocchi Bianchi)
There’s another problem: most companies are applying old change management playbooks to a technology that rewrites itself every week. Static governance frameworks and best practices that are often obsolete before the ink dries.
My take: the real competitive advantage of the next few years won’t be having access to AI or knowing how to use it. Everyone will have that.
It will be knowing how to turn individual knowledge into collective organizational capital — and doing it in a way that reflects what makes each company genuinely distinct.