AI agents managing 25-75% of finance and accounting tasks for many CFOs

While still a minority, a significant percent of finance leaders have offloaded at least half of their accounting and finance workload to agentic AI tools. 

Recent data from corporate finance solutions provider Maximor found that 27% of CFOs reported that between half to three-quarters of accounting and finance tasks are now managed by agentic AI tools. Most commonly, at 51%, leaders have automated between one-quarter to one-half of their workload. Meanwhile, 20% say they've automated up to a quarter of their workload and a tiny fraction, just 1%, say they've automated between 75% and 100%. 

"Partial automation is now standard. The real challenge is knowing which processes can run independently and which must escalate. Most tools still blur that line," said the report. 

Wanan - stock.adobe.com

The number of agentic tools they use, though, tends to be small. The poll found that while 88% of respondents said they use at least one agentic AI tool in their accounting and finance processes, the vast majority (84%) use just one or two. Only a small proportion of finance leaders use three or more such tools (just 4%); meanwhile, 12% said they don't use agentic AI at all. 

"Adoption is already underway, quietly and unevenly. The next gains will come from orchestration and control across tools, not from adding more of them," said the report. 

Many leaders (31%) believe that AI will eventually allow them to shave at least five days off the monthly close process (most believe it will be less than five), which also roughly lines up with the 36% who strongly agree that AI will allow them to spend more time on strategic work. 

But at the same time, faith that AI can deliver accurate results is uneven. Only about 14% of respondents completely trust the technology to deliver accurate data in accounting processes. A much higher proportion (41%) "mostly" trust that it will work, followed by 33% who "somewhat" trust it, 12% who "slightly trust" it, and 0% who said they do not trust it at all. This hesitation may be sound, as 86% of respondents have had one or more instances where their finance team encountered inaccurate or hallucinated data while using AI for finance tasks; of those, 49% said only once, and 38% said several times. 

This likely is why 67% of the survey respondents believe it is either extremely or very critical to have human oversight so as to ensure data accuracy in accounting. Conversely, 2% of the leaders polled said it was not critical at all. 

"Human oversight is not resistance. It is responsible adoption. CFOs want automation that knows when to act and when to pause for judgment," said the report. 

The data calls to mind other results indicating that people's embrace of AI could be outpacing their ability to properly oversee its implementation. A report last week from Deloitte, for example, found that a comfortable 74% majority of companies said they plan to deploy agentic AI sometime within two years, 23% are already using agentic AI at least moderately, and 85% of companies expect to customize agents to fit the unique needs of their business. But, at the same time, only 21% of companies say they currently have a mature model for governance of autonomous agents.