Leaders use AI mainly for routine tasks, survey finds
Tue, 19th May 2026 (Today)
General Assembly and EZRA have published a survey showing that business leaders still use artificial intelligence mainly for routine tasks. The study covered more than 500 leaders in the US and UK.
The findings suggest that, even as confidence in the technology grows, leadership use of AI remains focused on information searches, document summaries and email drafting. Searching for information was cited by 69% of respondents, summarising documents by 68%, and drafting emails by 58%.
More strategic uses were less common. Scenario planning and organisational design were each cited by 27% of leaders, while financial modelling stood at 28%.
The survey also found a marked gap between management tiers. Vice presidents trailed directors across several measures of adoption and confidence, including confidence in using AI tools without risking company data and the extent to which teams had reworked structures and workflows around AI.
Only 58% of vice presidents said they felt confident using AI tools without compromising company data, compared with 85% of directors. Just 39% said they had reworked team structures and workflows around AI, versus 71% of directors. Meanwhile, 68% said they did not know what vibe coding is.
Training effect
Training appears to be one of the clearest dividing lines in the results. Nearly four in five leaders, or 79%, said they had attended AI training or taken a course on AI, up from 69% in 2025 and 52% in 2024.
Two-thirds of respondents said their company offers leadership-specific AI training, and 88% of those with access said they had attended. Those who had taken that type of training reported broader, more advanced use of the technology.
Among leaders who attended leadership-specific AI training, 96% said their team uses AI tools regularly, compared with 82% overall. Some 88% said they understood how to use AI without compromising company data, versus 68% overall, and 97% said they had the knowledge to make AI-related vendor decisions, compared with 83% overall.
This group was also more likely to have a rubric for assessing good AI use and to feel able to judge employees' competence with the technology. By contrast, leaders without leadership-specific AI training were more likely to use AI for drafting emails and less likely to use it for qualitative data analysis or competitive and market intelligence.
They were also less likely to say they had successfully used vibe coding. Only 19% of those without the training said they had done so, compared with 49% of those who had attended.
"Leaders are getting more comfortable with AI, but comfort doesn't equal capability," said Daniele Grassi, Chief Executive Officer of General Assembly. "We often see that companies offering one-size-fits-all AI training have people at the top using AI the same way as interns-to write emails and summarize documents. Those who have received leadership-specific and role-specific AI training are much more likely to pursue transformational, strategic use cases."
Performance reviews
The research also points to a shift in how companies assess staff. Nearly half of leaders, or 47%, said AI use is now factored into employee performance reviews.
Among those organisations, 62% evaluate contributions to team-wide documentation or processes that use AI. Another 61% review automated reporting on AI tool use, 56% consider improvements in key performance indicators attributable to AI use, and 42% look at anecdotal evidence of how employees use AI to become more efficient.
Larger companies appear less likely to take that step. Only 34% of leaders at businesses with more than 10,000 employees said AI is a factor in reviews. Among companies that do not currently include AI in performance reviews, 54% of leaders said they oppose doing so.
Job concerns
The survey found that AI is already influencing hiring decisions. A third of leaders said they had eliminated a role or decided not to open one in the past year because they believed AI could do the job.
That share rose to 52% in software and technology and fell to 23% in professional services. At the same time, concern about AI's effect on leaders' own roles is increasing.
One in 10 respondents said AI is likely to replace their own job within a decade, and another 24% said it is possible. The share who said they do not believe AI will replace them within 10 years fell to 56%, down from 65% in 2024.
Broader adoption is still rising. Some 93% of leaders said they personally encourage their teams to use AI, up from 76% in 2025 and 60% in 2024, while 82% said their team uses AI tools regularly, up from 52% in 2024. More than half, or 52%, said they had reworked team structures and workflows around AI.
Yet barriers remain. Among leaders not using AI regularly, 49% cited security, privacy or regulatory concerns, 40% pointed to a lack of clear use cases, and 35% cited cultural resistance or a lack of leadership prioritisation. A third said they do not have the skills needed, almost double the 17% recorded a year earlier.
Nick Goldberg, Chief Executive Officer of EZRA, said the wider challenge for companies goes beyond access to AI tools. "As AI adoption scales, the differentiator isn't the technology itself, but the human skills leaders also need to develop around it," he said. "Organisations that thrive cultivate four critical leadership behaviors: curiosity to ask brave questions about value creation, discernment to protect quality, humility to adapt when answers aren't clear, and connection to share learning across the organization. Leadership development must also focus on these uniquely human skills that determine whether AI drives transformation or just incremental efficiency."