Survey shows top areas where AI is falling flat after employers replace people with new tech
More than a third (34%) of Canadian hiring managers who eliminated positions after implementing AI have since had to add those roles — or very similar ones — back, according to a survey.
Hiring managers cited a range of reasons for reinstating the positions, found the survey of 1,365 hiring managers, spanning finance and accounting, technology, marketing and creative, legal, administrative and customer support, and HR.
This tracks with other research, such as a 2025 survey by Orgvue that found over half of business leaders who have made AI-related redundances admit they got it wrong.
Of the 39% of business leaders who said they’ve made labour cuts because of AI, 55% said it was the wrong decision, according to the survey of over 1,000 senior business leaders and C-suite executives in several countries including Canada
‘AI correction hires’ in Canada
So what was the main reason for these “AI correction hires,” according to the new Robert Half survey? Hiring managers cited a range of reasons for reinstating the positions. The most common were that AI required more human oversight and quality control than expected, and that business demand increased, requiring more overall capacity — each cited by 38% of respondents.
Close behind, 37% said the eliminated roles involved relationship management that AI couldn't replicate, and an equal share said AI tools weren't fully or consistently adopted across their teams. A further 36% said the roles required institutional knowledge or context AI couldn't replace.
Productivity gains falling short of expectations were cited by 35% of respondents, as were risk or compliance concerns that emerged without a human in the role. A third of hiring managers (33%) said remaining team members experienced burnout or workload strain following the cuts.
A separate Robert Half survey of 1,005 Canadian workers found that when using generative AI, employees spend 33% of their total working time on a task checking accuracy and refining the final deliverable — pointing to an oversight burden that may not have been factored into early decisions to reduce headcount.
