The Research Question
In the field, Mustapha Benkalfate observes a paradox: leaders make major decisions about AI—investments, reorganizations, tool choices—while many remain at the stage psychology calls unconscious incompetence. His research aims to rigorously document these decision-making patterns, beyond consultant observations.
The Method
- A qualitative thematic analysis, seeking recurring patterns across different companies and sectors. A second, phenomenological study may then delve into the "why" behind these patterns.
- About fifteen decision-makers, half in similar contexts and half in varied contexts, to avoid limiting the study to a single profile.
- Companies with 1,000 to 10,000 employees: large enough to have different units and multiple levels of decision-making. He observes marked differences with structures of fewer than 100 people.
- Interviews of about one hour, recorded and transcribed under the university's protocol, until data saturation is reached. Participants provide consent and can request the results.
The Link to Deskilling
The research is linked to his other subject: the cognitive effects of generative AI. He has compiled a literature review on the "debts" created by uncontrolled AI use (cognitive debt, intention debt, epistemic debt), on deskilling, automation bias, and metacognitive laziness. It informs his keynotes and our pages on the studies and the glossary of cognitive risks.
A Consultant's Method Turned Research Material
Mustapha Benkalfate has been recording and transcribing his business meetings for years. For his July 2026 keynote in Nottingham, he analyzed these transcripts to highlight what leaders and recruiters actually call "AI fluency." He specifies that this is not an academic definition, but the sorted words of decision-makers. Read the keynote summary.
"What I share is based on what I have seen myself. I filter the research articles I read based on what I have actually been able to observe."
Why This Is Useful for Companies
The research provides a more solid foundation for Revenu IA's programs: executive committee training, use case diagnostics, and measuring the true maturity of teams with the AI Dynamic Assessment©.