New research conducted by Harvard Business Review Analytic Services in association with Nexthink, finds that four in ten (42%) survey respondents say their organization’s existing data/technology infrastructure doesn’t support AI monitoring. This could present a big barrier to understanding how AI is being used across the workforce.
The ‘Bridging the AI Acceleration Gap’ report is based on a global study of 259 respondents, all involved in their organization’s AI decisions, and finds little will change over the next 12 months. Only 19% of respondents say their organization will be implementing or enhancing visibility software during that period. Instead, the most common planned actions are AI education/training for employees (75%) and AI governance policies (67%).
“Employees are discovering, experimenting with, and integrating AI into their daily work long before many organizations have been able to articulate a clear AI strategy,” said Pedro Bados, CEO and co-founder of Nexthink. “By the time IT begins rolling out governance frameworks, operating models, and policies, it must consider software preferences, user behaviours, and ways of working that are already deeply embedded.”
That mismatch is already showing inside businesses. Over half (58%) agree that employees’ uptake of AI is outpacing their IT’s ability to monitor and govern use, while 45% agree that IT is being asked to govern AI without clear visibility into how it is actually being used.
The blind spot is biggest around informal (shadow) AI use, where employees use AI tools at their own discretion for tasks, instead of the tools their employer has defined for specific business purposes. While 46% say their organization has sufficient visibility into its formal AI use, only 27% say the same for informal AI use among employees. Nearly two-thirds (63%) say visibility into their organization’s informal AI activity should be improved. Security and/or data exposure is the leading concern about employees adopting AI at their own discretion, named by 79%.
“Visibility has become as important as capability,” added Bados. “AI is immensely powerful but to harness it effectively, executives need to be able to combine governance with AI innovation and automation with a deep understanding of real, day-to-day employees’ experiences in order to lead their organizations through this period of profound change.”
Methodology
The Harvard Business Review Analytic Services survey was conducted online between February and March 2026. It surveyed 259 members of the Harvard Business Review audience, all of whom are involved in their organization’s decisions about AI, and all of whom work at organizations already using AI to some degree. Respondents represented organizations with 500 or more employees across North America, Europe, Asia-Pacific, Latin America, the Middle East, and Africa.
