Few titles carry as much weight in balancing risk and opportunity as CFO. And that weight is only growing.
Over at least the last five years, and with mounting urgency post-pandemic, CFOs have faced a dual mandate: to modernize aging financial infrastructure and to safeguard enterprises against sophisticated fraud and compliance risks.
Artificial intelligence (AI) has become central to CFOs’ efforts, and not as a speculative experiment, but as a pragmatic tool to address acute vulnerabilities. For proof, look no further than the news this week from Anthropic and Deloitte, which have partnered to build AI compliance solutions; and Ramp, which added new AI agents to Bill Pay, its accounts payable automation platform; separately reveal.
The pattern emerging across industries is striking. Finance leaders are moving quickly to deploy AI where the business case is immediate and measurable, particularly in fraud detection, payments reconciliation and compliance. Yet these same leaders are proceeding deliberately, even conservatively, when it comes to applying AI across broader finance functions such as forecasting, capital allocation or strategic planning.
The result is a transformation that is both accelerated and constrained: The cautious rollout of AI in risk-sensitive areas contrasts with a longer arc of modernizing finance operations and replatforming legacy systems.
Risk Calculus
The latest PYMNTS Intelligence report, “From Experiment to Imperative: U.S. Product Leaders Bet on Gen AI,” captures this pivot well. Eighty-seven percent of product leaders now expect AI to improve fraud detection, 85% forecast better regulatory compliance, and 83% anticipate stronger data security.
“There’s a continuous evolution and … dynamic disruption in finance that requires CFOs to harness data and AI to make finance more efficient, more effective and substantially more strategic,” Raj Seshadri, chief commercial payments officer at Mastercard, told PYMNTS during a discussion for the B2B PYMNTS 2025 event, “B2B.AI: The Architecture of Intelligent Money Movement.”
“You can’t apply AI until you have really good-quality data at scale,” she added.
Against this backdrop, the first wave of AI adoption in corporate finance has largely targeted fraud and compliance because these areas offer both clear economic return and well-defined metrics for success. Sophisticated fraud schemes — from synthetic identities in consumer lending to shell-company payments in procurement — can evade rule-based systems. AI fraud detection models, by contrast, can surface subtle correlations among high-volume transactions and flag anomalies in near real time.
Even as CFOs focus on risk and compliance, many view AI as a gateway to rethinking the finance function’s strategic role. Cloud-native ERP systems and integrated data platforms are enabling more granular visibility into working capital, supply-chain performance, and customer behavior. AI applied to these datasets could unlock efficiencies and identify opportunities for growth.
Incremental but Irreversible
While AI promises faster and more accurate detection of irregularities, it also introduces new governance obligations.
According to data in the September 2025 edition of The CAIO Report from PYMNTS Intelligence, “How Agentic AI Went From Zero to CFO Test Runs in 90 Days,” none of the surveyed CFOs are willing to grant full, unfettered access to internal data and action permissions to agentic AI systems, and only a slim minority (8.3%) would allow moderate access.
At the same time, large multinationals have learned that, at times, retrofitting AI tools onto fragmented or partially digitized processes can produce more noise than insight. As a result, many CFOs are sequencing their AI deployments to align with broader system upgrades, fast-tracking high-impact fraud detection pilots while planning more ambitious analytics and forecasting initiatives for later phases.
The cautious pace in expanding AI beyond risk mitigation reflects a broader philosophy. Finance leaders are charged with stewarding capital and trust; they cannot afford experimentation that could destabilize financial reporting or introduce bias into strategic decisions.
But as the march of progress continues onward, some benefits are becoming harder and harder to ignore.
“We’ve been able to shave off a couple days in a month just by using AI tools to better reconcile data, better ensure that we’re having the right invoicing patterns,” Emanuel Pleitez, head of finance at Finix, said to PYMNTS in an interview posted Oct. 8.
“If you just start using AI today without needing to make the big five, 10% of your budget investment into it, you can actually extract and get five to up to 20% more productivity gains.”
Source: https://www.pymnts.com/