Everyone talks about generative AI, but in practice, only a minority use it every day to work, code or study. And among these intensive users, one profile comes up so frequently that it begins to intrigue researchers and HR alike.
In a column published in The Tribune and taken over by JVTECHthe work transformation expert Mélodie Ardouin puts forward a thesis that challenges the clichés. According to her, true early adopters of AI in business often share the same characteristic: an over-representation of neurodivergent profiles. What exactly are we talking about, and what does this change for the world of work?
Are early adopters of AI really mostly neurodivergent?
Mélodie Ardouin states it bluntly: “These early adopters are not exceptions. They are not all eccentric tech profiles or geeks from the start either. They share a common point that few organizations identify: an over-representation of neurodivergent profiles. »
By neurodivergent, we are talking in particular about people with ADHDautism spectrum disorders, Dys disorders or other forms of neuroatypia. Their common point: a high cognitive load on a daily basis. Organizing, prioritizing, launching a task or keeping track of a project requires more energy than average. According to the study Global Neuroinclusion at Work ofEY (2025), these professionals are 55% more likely to use AI every day than their colleagues.
Why is AI becoming a cognitive assistance for neurodivergent profiles?
For many employees, ChatGPT, Claude Or Gemini mainly to save time. For certain neurodivergent profiles, the tool goes further: it acts as a real cognitive assistance. The AI accepts messy thinking, incomplete requests, then proposes a plan, a list of actions, a structured email.
Concretely, an employee with ADHD can ask the AI to break a huge project into small steps, prepare a report from chaotic notes or rephrase a professional message. Where traditional tools require initial rigor, AI adapts to the user’s brain. Companies like JPMorgan Chase report 90 to 140% higher productivity on certain engineering tasks for their neurodivergent employees, while Enabled Intelligence50% of whose workforce is neurodivergent, achieves 97% detection accuracy compared to 70% for the industry average.
Businesses: What to do if your best AI users are neurodivergent?
This is where the paradox becomes brutal. The joint study EY – Microsoft (2024) shows that 76% of neurodivergent employees say they perform better at work thanks to AI, particularly in terms of communication, memory and concentration. However, only 25% of them truly feel included in their organization, 91% say they face career progression barriers and nearly 40% plan to leave their position within the year.
In other words, companies are investing heavily in suites of AI tools without seeing that their best beta testers are already there, often on the sidelines. Identifying these pioneers, recognizing their role as pioneers and adapting the work environment to the plurality of brains becomes a strategic issue. For many organizations, the real question is no longer having an AI strategy, but knowing how to listen to those who already use it as a cognitive prosthesis on a daily basis.