Research Reports
-
Most research on AI comes from vendors, consultancies and surveys of executives. These reports ask working professionals what they do and what they see, and they compare the answers between groups. The recurring finding is a gap: between leaders and users, between belief and evidence, between what organizations expect and what happens. Each report names the gap it found and explains the behavioral mechanism behind it.
Each report presents findings from a survey Behavieural designed and analyzed. They test what behavioral science predicts about AI adoption, confidence, and value, and they show where common assumptions break down.
-
The Visibility Gap
1 November 2026
Why organizations believe AI is creating value but cannot show it.
The Visibility Gap asks whether people believe AI is producing value, whether anyone has a way to verify that belief, and whether the people using AI and the people overseeing it see the same thing. It surveyed 467 working professionals in organizations that use AI. Two in three (67%) agree that AI is producing meaningful value, yet only one in five (21%) report that their organization has a formal method for measuring it, and 31% do not know whether one exists.
The report also shows that visibility depends on where a person sits. Individual contributors are more than twice as likely as managers and executives to say they do not know whether AI value is measured (42% against 17%), and fewer can see how AI use connects to outcomes that matter. The report introduces the idea of leakage, a reliable gap between perceived time saved and perceived redeployment of that time, and sets out four steps that separate organizations that can show what AI returns from those that rely on belief.
-
The Adoption Gap
5 October 2026
What actually predicts whether people use AI?
The Adoption Gap tests a common assumption: that stalled AI adoption is mainly a technical or change-management problem. The survey asked 523 working professionals in healthcare, finance and insurance, and manufacturing and operations about their use of AI tools, their intent to keep using them, and the conditions around them, including technical quality, incentives and culture, regulatory concerns, change management, and a set of psychological factors.
The psychological factors, taken together, were the strongest predictor of usage and intent to continue, with an effect roughly twice that of technical quality or incentives. This held within each of the three sectors. Conventional change-management measures did not significantly predict adoption. The report is careful about scope: the finding concerns the psychological factors combined, not any single factor, and the study is a cross-sectional survey, so it shows association and not cause.
-
The Confidence Gap
7 September 2026
Do the people who oversee AI and the people who use it see the same thing?
The Confidence Gap examines how professionals involved in evaluating and overseeing AI judge its trustworthiness, and how those judgments differ between individual contributors and managers and executives. The survey covers healthcare, finance and insurance, and manufacturing and operations, with a working sample of 770 respondents.
The central finding is that confidence in AI is not evenly shared inside an organization. Leaders and the people closest to the work tend to hold different views of how well AI is performing and how safe it is to rely on, and the report shows where those views diverge. It also maps its findings to the engagements most relevant to each pattern.