3.3times
The most skilled share more
Among the 25% of respondents most at ease with prompting, highly helpful profiles are 3.3 times as common as individualist ones.
33 highly helpful10 individualists
AI at Work Barometer 2026 edition
Each year with the Worklab, we measure how professionals use generative AI and what it changes for their teams. Answer in 7 minutes and see how your profile compares with the 2025 respondents.
The questionnaire is in French.
Survey open until December 31, 2026, with respondents so far
A study run with 
219respondents to the 2025 edition
15industries
3paradoxes uncovered
Prompting skills get shared where the hierarchy allows it
The central finding of the 2025 edition, based on 219 respondentsThe more hierarchical the organization the less skills spread
-
Low hierarchyr = 0.52
The better people are at prompting, the more they help their colleagues. Skill benefits the whole team.
-
Moderate hierarchyr = 0.26
The link is still there, half as strong.
-
High hierarchyr = 0.14
The link almost disappears. The gains from AI stay individual.
Link between prompting skill and helping colleagues
0.52strong link between prompting and helping
Three more findings against the grain
They cover helping colleagues, caution with data and the framework the company sets. All three challenge what people often say about AI at work.
5.28out of 7
The most cautious handle the tool best
People who deliberately limit what they share with AI get a higher prompting score. Care with data goes hand in hand with skill.
Average prompting skill score, on a scale of 1 to 7.
8.2%
Shadow AI grows where rules are missing
8.2% of respondents use AI often, with little trust in their organization and no resources from it. Where the company equips its people, trust follows.
One square per 1% of respondents.
0.65weight of prompting skill in how often people use AI, three times that of organizational support (0.19)
43%of respondents used generative AI every day in 2025
0.37correlation between the resources the company provides and trust in how it manages risks
What to take away for your AI projects
What the 2025 results tell executives and managers who are bringing generative AI into their teams.
Train people in prompting first
Prompting skill weighs three times more than organizational support in how often people use AI. The fastest lever is short, job-specific training with concrete use cases and a measurement before and after.
Make room for sharing
Skill fuels helping where the hierarchy lets ideas flow. Set up sharing rituals and recognize the time spent helping a colleague: your AI champions then become go-to people for the whole team.
Equip your teams to reduce shadow AI
Unsanctioned use grows out of a gap: few tools provided and vague rules. When the company provides tools, training and data rules, its teams' trust follows.
Rely on your most cautious people
The people who limit what they share with AI are also the most skilled. Bring them in to write your AI policy: governance and performance move forward together.
AI amplifies what your organization already allows.
Your profile in 7 minutes
The 2026 edition repeats the 2025 questions to track year-over-year trends, and adds eight questions about what has changed since.
Tools usedAgents and automationsTraining receivedCompany policy
- 1You answerOne question per screen, no account or email required. Your answers stay anonymous.
- 2You see your profileHow you prompt and how you help colleagues, compared with the 219 respondents from 2025.
- 3You get the resultsIn late January 2027, if you leave your email at the end.
The questionnaire is in French.
Sample profile
Your profile
Collaborative champion
You're skilled at prompting and your colleagues benefit from it.
You describe a workplace with little hierarchy.
Two perspectives behind the study
One looks at how people use AI, the other at how people work together. They meet in every edition, from designing the questionnaire to reading the results.
As founder of Jessy Martin Consulting, he helps marketing and communications teams use generative AI, from strategy to everyday tools.
As founder of Jessy Martin Academy, a Qualiopi-certified training organization [French quality certification for training providers], he trains professionals in Claude and in using AI at work.
More about Jessy
Sacha Lopez
Founder of the Worklab
Founded in 2015 in Nantes, the Worklab is a consulting and training firm specializing in collaborative work and new ways of working: facilitation, training and advisory in collective intelligence. It is a founding member of Les Artisans du Changement.
Sacha Lopez is the author of Guide de survie aux réunions [a survival guide to meetings] and Guide de survie aux réunions à distance [the same for remote meetings].
worklab.fr (in French)“Jessy is a true expert with a passion for AI and innovation. I was lucky enough to work with him on running large professional events, and he proved to be both a great teacher and a very good facilitator.”
- 20 identical questions from one edition to the next, on a scale of 1 to 7
- 4 dimensions: prompting, helping, hierarchy and support, privacy
- Correlations, moderation tests and regression for the analysis
Your questions
Anyone who already uses generative AI tools at work, whatever their job, industry or company size.
Yes. The survey does not record your name, email or IP address. If you leave your email to receive the results, it is stored separately and never linked to your answers.
About 7 minutes. Your profile appears as soon as you answer the last question.
Data collection closes on December 31, 2026. The results will be published on this page in late January 2027.
Join the 2026 respondents
The more people take part, the more accurately the results will describe what happens inside teams. You see your profile as soon as you finish the questionnaire.
Answer in 7 minutesThe questionnaire is in French.
Survey open until December 31, 2026, with respondents so far