Dive Brief:
- CFOs did not see artificial intelligence meaningfully affecting employment at their organizations last year and don’t anticipate a large effect on headcount from the technology over the course of the current year — but are expecting a change in the overall skill makeup of their workforces, according to a working paper published this month by the National Bureau of Economic Research.
- Finance chiefs expect the introduction of AI will lead to a decline in the number of “routine clerical roles” over the next three years, forecasting routine clerical employment to slump by 2 percentage points by 2028. Additionally, the share of technical roles, including engineers, data scientists and analysts, is expected to grow during that period — rising by 0.62% in 2026 and 1.35% by 2028, the survey of 750 CFOs conducted by Federal Reserve Banks of Atlanta and Richmond economists found.
- Notably, however, the paper also found lowering costs — including labor as well as non-labor expenses — to be one of the least important motives for AI investments among firms. “Overall, therefore, in the near-term, the goal of investing in AI is not to reduce workforce or costs but rather to improve productivity,” the research found.
Dive Insight:
The study examined CFOs’ responses concerning how AI use has helped to improve worker productivity, asking finance chiefs to note a percentage change over the past year and the next two years.
For 2025, the mean labor productivity growth attributable to AI was 1.8%, the survey found. Over the course of this year, productivity gains are expected to “grow considerably,” with mean reported labor productivity expected to reach 3%.
“Labor productivity gains are positive, vary across sectors, and are expected to strengthen in 2026, with the largest effects concentrated in high-skill services and finance,” according to the research, which surveyed four sectors, including finance, manufacturing and construction and high-skill and low-skill services.
The company also asked CFOs to report a percentage change in revenue and employment due to their companies’ use of AI and found that these “implied productivity gains” were also positive, with finance seeing implied labor productivity growth of about 0.8%.
The findings come as focus on AI skills, particularly within finance, is beginning to grow among employers and CFOs: A recent study by Datarails, for example, found approximately one in three finance job postings now explicitly reference AI skills, compared to one in four just a year prior. Finding effective ways to implement AI inside of the finance team itself is also becoming a top priority for many CFOs: Hewlett Packard Enterprise finance chief Marie Myers told CFO Dive last month that her finance team looks to do more with agentic AI tools.
Finance chiefs also appear to have a positive view of AI’s ability to bolster worker productivity in the future, though gains may not appear in the short-term, according to the survey. The study did identify what it terms as an AI “productivity paradox,” noting survey respondents consistently identified larger AI productivity gains “than those implied by contemporaneous changes in revenue and employment.”
However, the research also noted that this gap is likely due to “delayed output realization and quality improvements that are not yet captured in measured revenues.
“More broadly, firms’ conceptual notion of productivity appears to extend beyond mechanical revenue-per-worker calculations, encompassing improvements in workflows, task efficiency, and organizational capacity whose revenue effects materialize only gradually,” the paper noted.