Dive Brief:
- The share of U.S. enterprise leaders reporting adoption of artificial intelligence agents in their finance departments rose to 41% in 2026 from 24% in 2025, according to a survey by workflow automation company Zapier.
- Finance recorded the largest year-over-year increase in AI-agent adoption among the corporate functions tracked, with its 17-percentage-point jump. Human resources followed, rising from 30% to 43%, while operations climbed from 47% to 59% and sales increased from 26% to 37%.
- The findings suggest finance teams are becoming more comfortable deploying AI agents after initial concerns about accuracy caused them to lag behind, Zapier CFO Ryan Roccon said in an interview. “The models that existed a year ago weren't always accurate or reliable enough for finance, particularly when it came to financial analysis or deeper data extraction,” he said.
Dive Insight:
Despite the adoption surge, concerns remain. Across all respondents, 27% cited data privacy, security and regulatory compliance risks as barriers to adopting AI agents, while 24% pointed to accuracy, reliability and errors.
Only 30% of leaders at organizations using AI agents said their agents complete more than 60% of assigned tasks without human intervention, according to the survey.
Roccon dismissed the idea that recent reports of rogue AI models should cause CFOs to reconsider or pause adoption. He warned that companies that shy away from the technology risk putting themselves “at a major disadvantage.”
“If you are concerned about these models and the risks that they present, start with things that are low risk, that are contained,” Roccon said.
Zapier's survey found 71% of organizations have AI agents in production, up from 40% in 2025. Nine in 10 respondents said their organizations were certain or likely to increase investment in AI agents over the next year, compared with 84% in the 2025 survey.
Document analysis and summarization, along with data extraction and entry, were the most common uses of AI agents overall, according to the survey.
The survey, conducted from Sept. 16-21, included 513 U.S. business leaders.
Finance was the least likely corporate function to use AI agents in last year's survey. Roccon attributed the increase to improvements in model accuracy, as well as better integration with tools such as Microsoft Excel and Google Sheets.
Roccon said accounting is a good starting point for finance teams adopting AI because many processes are rules-based and their outputs can be verified.
At Zapier, he said, many month-end processes run through an AI system, but a human must approve transactions before they are posted to the general ledger.
“Nothing hits the general ledger ever without a human clicking approved,” he said.