UK tech leaders plan to hire AI managers within a year

More than half of UK technology leaders plan to hire a dedicated AI manager within the next year, according to research from Asana. The findings suggest a growing push to give formal oversight to AI agents as they take on more work inside organisations.

Among UK leaders whose organisations do not already have a dedicated AI-agent manager, 14% said they were actively hiring for the role and 54% planned to do so within 12 months. The research also found that 57% of UK leaders said their organisation already had a dedicated AI-agent manager, compared with 33% of UK workers, suggesting a gap in how visible those responsibilities are across the workforce.

The survey suggests employers are moving beyond testing AI tools and are now grappling with how to supervise them. As agents are used for a wider range of tasks, businesses are increasingly deciding who should monitor performance, review mistakes and control access to data.

Both leaders and workers ranked performance monitoring as the most important responsibility for an AI-agent manager, cited by 61% of leaders and 55% of workers. Reviewing incidents and mistakes came next, selected by 55% of leaders and 47% of workers. Managing permissions and data access followed, named by 49% of leaders and 46% of workers.

The findings also show that supervising AI agents is taking up a notable share of working time. In the UK, leaders estimated they spend 36.49% of their time managing agents, compared with 30% for workers. The figures were higher in the US, where leaders put the share at 49.4% and workers at 42.1%.

Coordination burden

Alongside the push to create management roles, the research found that employees are dealing with practical problems as they use multiple agents at work. Among UK workers who use two or more AI agents in a typical week, 40% said switching between tools, deciding which to use, managing outputs or resolving conflicting results took up more than 10% of their working time.

That points to a less straightforward picture than the productivity claims often attached to AI deployment. While organisations are introducing agents to help staff complete work, some employees report spending part of their day coordinating the tools rather than reducing their own workload.

Overlap is another issue. Among UK workers whose organisations use two or more agents, 33% estimated that more than a quarter of the work completed by agents was duplicated, repeated or substantially overlapped with work carried out by another agent.

The research also found that many workers do not think the tools are focused on their most pressing needs. Only 18% of surveyed UK workers said the agents being introduced addressed the tasks and problems employees most needed help with. A further 33% said agents addressed some important needs but left significant gaps, while 21% said they mostly addressed lower-priority or more convenient tasks.

Autonomy and control

Questions about oversight are tied to wider concerns about how much freedom AI agents should have inside organisations. Security and privacy were the most commonly cited barriers to scaling their use, named by 32% of UK leaders and 28% of UK workers.

Most UK leaders said they would be comfortable allowing an agent to access information and complete actions without reviewing every step. The share stood at 78% among leaders, compared with 56% among workers. Only 16% of workers said they would be very comfortable with that level of autonomy.

Workers also highlighted specific concerns about access and accountability. Some 28% said they worried that agents might access sensitive or confidential information without people realising. Another 23% were concerned that access permissions could expand without sufficient review, and 21% worried that an agent might act on information it should not have accessed.

The data suggests companies face a management challenge as much as a technical one. Deciding what an agent can access, which tasks it should handle and when a person should intervene appears to be becoming a formal operational question rather than an informal responsibility shared across teams.

Use of multiple AI agents is already more common among leaders than workers. The survey found that 66% of UK leaders personally used or worked with three or more agents in a typical week, compared with 29% of UK workers.

That difference may help explain why management structures are emerging first among senior technology teams. As agent use spreads, the gap between leadership expectations and worker experience may shape how quickly dedicated AI manager roles become standard across UK businesses.

The research was conducted among 4,502 respondents in the UK and US, including 1,501 CIOs and AI leaders and 3,001 workers. The UK findings drew on responses from 2,252 CIOs, AI leaders and AI-using office workers.