First who, then what: AI only transforms organizations ready to welcome it.

What is the pitfall that you most often observe among your clients? Many companies purchase licenses, deploy tools, or launch top-down training plans with the best of intentions. However, when these initiatives are disconnected from the daily lives of teams, adoption does not follow.

Believing that the tool or information is enough to create use is an illusion. MIT measured this in its study The GenAI Divide published in 2025: 95% of the organizations studied were not getting any measurable return from their generative AI projects. Researchers focus less on the quality of the models than on their discrepancy with daily operations.

Hence your conviction: “First who, then what”? Exactly, it’s a conviction that has driven Extia since our beginnings and to which AI is now giving new relevance. Making AI a performance lever requires combining team commitment and technological mastery.

Another striking figure is the one shared by BCG in 2025* in its annual “AI at Work” survey: the share of employees positive about AI increases from 15% to 55% when their hierarchy is strongly involved. However, today, only 1 in 4 field employees benefit from this support. Starting with the “who” is also a question of profitability.

This is about avoiding wasting time, budget and team energy on projects that will ultimately never be adopted. What do you recommend to promote team engagement? Place the profession at the heart of the reflection.

Ideation and prioritization of projects must come from the teams: they are the ones who know where AI will be useful. Then, question before automating. AI invites us to be discerning about our intentions and needs.

Automating a worthless task is once again wasting company resources. This is an opportunity to fight against the mille-feuille effect, one of the main risks linked to AI from an organizational and structural point of view. Third point: being able to do more, or faster, is not an invitation to quantity, but a reminder of the virtues of quality.

The way in which freed up time is reinvested says a lot about team engagement: relationships between colleagues, innovation, improvement of products and services, or simply time to breathe. Finally, if there is a key element that conditions the success of AI acculturation programs, it is trust. Without it, adoption remains confined to individual gains, sometimes hidden.

The trend of “Bring Your Own AI”, or shadow AI which continues strongly today, is a strong signal. Employees want to move forward, it is up to the company to create a framework where uses are shared openly. At the same time, how can we move from experimentation to results?

First, set the course. This involves setting a clear trajectory and prioritizing use cases with high added value. Second, establish organizational governance.

Machine-translated from French