It started with moving from legal pads to spreadsheets. Then shifted from spreadsheets to accounting and payments automation. Next, gen AI came along and the trend from manual finance management and forecasting took on a new speed and urgency. Now a new, more sophisticated and complicated form of AI is here, one that requires CFOs to let go of the wheel in a big way.
It’s agentic AI. It promises to turn yesterday’s month-end close into a continuous, autonomous, automated predictive engine. On the consumer side, agents are being pitched as a shopping assistant and task manager. But the more impactful usage will come from the business side of the economy. With capital markets bracing for tighter spreads and relentless rate volatility, the pressure to capture AI’s agentic speed advantage is suddenly urgent. CFOs know that, but …
The same executives who champion agentic AI’s autonomy are privately questioning whether the technology is battle-ready. PYMNTS Intelligence’s coverage—including last week’s (June 21) interview with Trustly chief legal officer Kathryn McCall—underscores a growing fear that agentic AI can go “off script,” exposing firms to cascading payment errors, sanctions screening failures and headline-grabbing compliance fines. Infrastructure gaps, large language model transparency and still unsettled audit and regulatory standards mean accountability remains a moving target. For finance leaders, the message is stark: Harness AI’s momentum now, but build the guardrails before the next quarterly call—or risk owning the fallout.
It’s no surprise that enterprise CFOs aren’t ignoring agentic AI, but neither are they racing to deploy it. Familiarity with the technology is nearly universal, as expected from the hype. But the execution-level confidence isn’t there yet. Just 15% of executives surveyed by PYMNTS Intelligence are considering putting it to work, and most of those remain in the early evaluation stage. The hesitation comes even as the use of gen AI has surged. Among its use cases: creating content, fielding customer service queries, coding software and analyzing data. Since March 2024, the share of CFOs deploying the technology to innovate products and services has jumped 21%. Meanwhile, its use for spotting fraud and errors has climbed 31% over that same period.
Here’s another interesting finding: CFOs are more likely to trust agentic if they’ve been successful with gen AI. The momentum behind agentic AI, which goes one step further than its precursor by autonomously making decisions and carrying out tasks independently of humans, builds at companies where the latter has lived up to its promise. Among highly automated firms, 67% are weighing agentic AI adoption, compared to just 11% of their low-automation peers.
Likewise, firms reporting strong return on investment (ROI) from gen AI are twice as likely to increase their spending compared to low-automation CFOs. Among the big questions: Will the technology take away jobs? Across the board, CFOs say that gen AI has reshaped staffing needs. All report needing more analytically skilled workers, and nearly all agree that the overall skill mix across teams has changed. AI is on its way to presenting another level of automation to CFOs, one that will require a prerequisite course in gen AI deployment to understand and a more cautious, patient approach.
These are just some of the findings detailed in “The CAIO Report: The Agentic Trust Gap: Enterprise CFOs Push Pause on Agentic AI,” a PYMNTS Intelligence and AI-ID collaboration. This edition examines the future of autonomous AI in enterprise settings and draws on insights from a survey of 60 CFOs conducted in May 2025.
Agentic AI Adoption Lags Despite High Familiarity
CFOs know the term, but not the tech.
Technology companies developing and selling agentic AI tools have done a strong job of making sure companies know what’s on offer. Every CFO now knows about the technology and its potential to fully automate a host of labor-intensive business functions, from cybersecurity to product development. The reality is that very few companies are actually using it.
Despite nearly universal familiarity, just 15% of CFOs are even considering putting agentic AI to work in their companies—and 85% say they have no current plans. Among those considering climbing on board, 89% remain in the evaluation phase, with no concrete roadmap in place.
But that top-line number obscures a more nuanced story. Adoption interest closely tracks how far firms have already gone with gen AI automation. More than two in three CFOs from high-automation enterprises are considering agentic AI. That figure plunges to 11% among firms still dependent on manual workflows. Firms that already automate processes more are more willing to explore what comes next.
The same split appears by self-reported impact. One-quarter of high-impact CFOs—those whose companies are most active in using gen AI—are at least considering the agentic version. In medium- and low-impact roles, the share drops to 12% and 7%, respectively. The gap reflects how closely interest in autonomous systems aligns with C-suite support.
Even among executives considering adoption, active test runs of the technology remain rare. Just 6% of high-impact CFOs, meaning executives at highly automated firms, say they’re exploring adoption within the next year.
Gen AI Use Cases Surge While Investment Cools
Agentic AI remains an idea, not a project.
What is clear is that companies are not accelerating their use of agentic tools. Among CFOs considering adopting the technology, 89% say their organizations are still evaluating its relevance and fit. Just 11% are running pilots or proof-of-concept trials. Another 11% report active vendor partnerships. The overlap means that very few are doing more than research and actively using the technology.
Even among the most engaged segment—high-impact CFOs—activity remains limited. One in five say their firms are piloting agentic AI. The same share reports working with vendors to explore early use cases. That leaves four in five still in the planning phase. These CFOs are best positioned to move forward—but most have yet to do so.
Engagement is nonexistent among lower-impact personas. No medium- or low-impact CFOs say their firms are evaluating agentic AI. None are testing deployments. None are working with partners. This total lack of movement reinforces the idea that strategic influence is a precondition for even limited adoption steps.
Caution among more uncertain firms may reflect broader pressures. Gen AI has become essential—90% of CFOs say it’s now critical for financial planning. Use is expanding across fraud detection, service innovation and real-time query response. Still, investment in the technology is slowing. Just 27% of firms are increasing budgets this year—down from 53% last year. That contraction likely spills over to agentic AI, which requires even greater confidence and trust in the technology.
Until firms commit time, budgets and cross-functional leadership, agentic AI could sit stuck where it is today: as a widely known tool with promise for transforming business operations, but no timeline for that to happen.
Automation and ROI Drive Future Adoption Plans
Over the past year, CFOs have doubled the importance they place on Gen AI.
While companies have been slow to deploy agentic AI, their use of gen AI has nearly doubled. Nine in 10 CFOs now say it’s very or extremely important to financial planning and analysis—up from just 47% in March 2024. That shift reflects not just broader use but growing operational dependence on the technology. Gen AI has become core to how CFOs plan, monitor and execute across financial domains.
Usage continues to expand. Since March 2024, gen AI’s use for product and service innovation has grown 21%. Deployment for spotting fraud and operational errors is up 31%. Three out of four CFOs now say they use the technology to generate real-time, automated responses to customer queries.
It all demonstrates that gen AI is no longer siloed to a narrow, individual task: CFOs say the technology’s value reaches across core functions. The tool ranks as highly important for financial reporting (87%), cost management (85%) and capital management (88%). More than 70% say it’s critical for strategic planning, risk management and working capital optimization. The shift isn’t limited to back-end efficiency—it spans decision making, compliance and customer experience.
Still, enthusiasm is now tempered by caution. Just 27% of firms are increasing their gen AI budgets in 2025—down from 53% a year ago. Investment plans now track closely with results. Among firms reporting strong ROI, half are increasing spend. Among firms with negligible ROI, that number drops to 17%.
Automation levels matter, too. Highly automated firms are three times more likely to consider agentic AI adoption than low-automation peers. These firms have already laid the groundwork—and now view agentic AI as a logical next step. Firms that have scaled gen AI—and benefited—are the ones continuing to invest.
Gen AI Shifts Workforce Needs and Skill Priorities
CFOs agree they need more analytically skilled workers in their businesses.
Gen AI has started reshaping the talent CFOs need to transform how they operate. All CFOs now say implementation of the technology has increased their need for more analytically skilled workers. In March 2024, that share stood closer to 60%. The shift in the past year reflects growing hands-on use and a rising urgency to close capability gaps.
The impact goes beyond analytics alone. Ninety-seven percent of CFOs say the overall mix of skills required across the business has changed. As gen AI automates workflows and elevates data as a core asset, all with the involvement of humans, CFOs increasingly need talent that can evaluate, interpret and act on machine-generated output.
Tech teams are expanding in response. Eighty-three percent of CFOs now say they need a larger technology or IT support team. That’s up 25 percentage points from 58% in March 2024. Demand for high-skill tech roles has grown as gen AI tools become more embedded in daily operations.
At the same time, demand for lower-skilled labor is falling. More than six in 10 CFOs say their need for such workers has declined. Nearly half (45%) say they no longer need to hire as many employees as possible in certain businesses. The trend points to gen AI’s role in reducing headcount pressure while reallocating hiring needs to higher-value, higher-skilled roles.
The workforce shift is structural. Skills once considered specialized—like data fluency and systems thinking—are fast becoming baseline requirements. gen AI isn’t just changing the tools teams use. It’s changing who organizations need to hire, train and retain to stay competitive.
Read More
PYMNTS Intelligence is the leading provider of information on the trends driving AI adoption patterns. To stay up to date, subscribe to our newsletters and read our in-depth reports.
Methodology
“The Agentic Trust Gap: Enterprise CFOs Push Pause on Agentic AI” is based on a survey of 60 enterprise CFOs conducted between May 6, 2025, and May 16, 2025. The report examines how CFOs are evaluating and adopting agentic AI as a follow-on to their existing gen AI investments. Respondents represent U.S.-based companies with $250 million or more in annual revenue across a range of industries. The sample included CFOs with varying degrees of strategic influence and operational impact, enabling analysis by automation maturity, gen AI return on investment (ROI) and role-based authority. PYMNTS Intelligence designed the survey instrument to measure both adoption intent and current deployment patterns, with an additional focus on budget planning, staffing and enterprise capability gaps.
