Dive Brief:
- Finance teams are spending an average 26% of their work time verifying or correcting AI-generated finance outputs, underscoring the limits of the technology even as its use becomes widespread, according to a Datarails report released Tuesday.
- The survey of 270 CFOs and finance leaders found that 96% spend at least 10% of their work time checking or correcting AI outputs. Eight percent spend more than half their work time on such reviews, Datarails said.
- “What this survey shows is that although AI adoption is now standard for CFOs, caution hasn't disappeared,” Datarails CEO Didi Gurfinkel said in a statement.
Dive Insight:
The findings come as finance organizations continue to expand their use of AI. More than half (53%) of respondents plan to expand their AI licenses over the next 12 months, while 32% said their organizations exceeded their AI budgets by at least 10% over the past year.
Lack of auditability is the biggest obstacle to greater trust in AI, with 75% of respondents citing it as a reason they hesitate to use the technology for mission-critical finance tasks. Concerns about accuracy and hallucinations followed at 71%, while 54% cited regulatory or compliance concerns.
“Significant hurdles remain, especially around trust, governance, and auditability, as well as ongoing challenges related to scattered and siloed data,” the report said.
At the same time, Datarails found 76% of respondents face high or very high pressure to fully implement AI, but only seven percent said their finance function is fully ready to implement the technology across all workflows.
The survey also found 60% of respondents are redeploying employees to higher-value work as AI takes on more routine finance tasks. Only three percent said they are cutting finance headcount because AI has replaced work.
Datarails said respondents were surveyed in July and worked at U.S. organizations with more than 1,000 employees and at least $100 million in annual revenue.
