With artificial intelligence becoming a more important tool for businesses, a new report from Datarails says finance teams are spending more than a quarter of their working week (26%) verifying or correcting output from AI tools.
Datarails, an AI finance operating system for the office of the CFO, surveyed 270 chief financial executives—all of whom said they use AI in some capacity for finance processes. Because of that, CFOs are facing a “verification burden”: 96% spend at least 10% of their work time verifying or correcting finance-specific AI outputs, according to Datarails’ 2026 CFO Sentiments Survey: How AI is Changing Finance Departments report. Another 8% spend more than half of their work time on such checks.
With AI use intensifying, spend challenges have surfaced: almost one-third (32%) of CFOs report that their organizations exceeded their AI budgets by at least 10% in the past 12 months, the report says. At the same time, more than half (53%) of CFOs are planning to expand AI licenses across their organizations over the next 12 months.
“What this survey shows is that although AI adoption is now standard for CFOs, caution hasn’t disappeared,” Didi Gurfinkel, CEO and co-founder of Datarails, said in a statement. “Finance teams have stopped asking ‘will we use AI?’ and started asking ‘how am I going to most effectively check its work?’”
Auditability emerges as a top challenge
Overall, lack of auditability (75%) is the biggest reason for CFOs’ hesitation in fully trusting AI tools with mission-critical finance tasks, followed by concerns about accuracy and hallucinations (71%) and regulatory or compliance concerns (54%) regarding AI outputs.
Nearly two-thirds (65%) of CFOs say their most common AI frustration is outputs giving confident answers based on the wrong data, followed by a situation experienced by 56% of CFOs where team members have been given materially different outputs from a large language model despite using the same prompt and data. The report points out that unless an LLM is restricted to governed, consolidated, contextualized data, successive responses frequently differ.
The survey finds that just 5% of finance leaders currently trust AI to produce board-ready financial reports without human review, and only 4% trust it with their month-end close.
In a parallel finding, a new technology category is emerging: 32% of finance leaders are prioritizing a “finance operating system”—a governed data layer built for AI. In terms of priorities, this new category is second only to traditional planning and financial planning and analysis tools (42%).
AI-driven layoff fears are unrealized
Some good news: 60% of CFOs say that as AI accomplishes more everyday finance tasks, they are actually redeploying staff to higher-value work rather than reducing headcount. Despite fears around job losses from AI, only 3% of chief financial executives are actively cutting full-time equivalent employees in the CFO’s office due to AI.
“Fears of job losses have been largely allayed, but the challenge of AI output verification is critical,” Gurfinkel said. “Tackling it will allow finance teams to spend less time doing intensive checking and let them focus instead on the strategic work they were actually hired to do.”
Only 4% have a ‘single source of truth’
Despite decades of touted progress, only 4% of organizations have achieved a “single source of truth” for their finance and operational data. Overall, 73% of CFOs say their data is “mostly centralized,” while 23% rely on disconnected systems and manual reconciliation. Teams whose top operational challenges are manual reporting and data consolidation are also the most likely to report AI generating a confident answer based on the wrong data (86% vs. 65% of all respondents), the report says.
Meanwhile, more than three-quarters (76%) of CFOs are under high or very high pressure to fully implement AI, but only 7% say their finance function is fully ready to implement it across all workflows.
Additional findings
Other key takeaways from the Datarails report include:
- Copilot frontrunner as main tool in the CFO’s office: Finance teams are running an average of 2.5 general-purpose LLMs, with Microsoft Copilot (93%), ChatGPT (65%), and Claude (64%) the most widely used—often in tandem.
- More than one in eight AI budgets have no owner: 84% of CFOs own their organizational AI budget, but 13% report no clear owner at all.
- Spending on the rise despite the trust gap: 53% of finance leaders plan to expand their AI licenses over the next 12 months; only 7% plan to cut or consolidate AI tools.
Photo credit: Unsplash
Sign in to get access to this free resource, and all of our whitepapers and reports.
Download this content today!
Register Now Already registered? Click here to Log In
