Finance Leaders are Feeling the Pressure to Adopt AI

Finance leaders are feeling the pressure to adopt Artificial Intelligence  (AI) because boards, investors, and competitors increasingly expect finance teams to use it to improve efficiency, reduce costs, and deliver faster insights. While AI can automate processes and enhance decision-making, many CFOs are balancing these opportunities against concerns around governance, compliance, data quality, and proving return on investment.

AI has moved out of the pilot phase and into the day-to-day operations of finance departments everywhere. Tax, compliance, financial close, accounts payable, and invoicing are all being reshaped by autonomous AI agents, and the finance leaders overseeing these functions are under real pressure to adopt AI quickly, often before the systems needed to manage it responsibly are fully in place.

A recent survey of more than 1,500 CFOs and senior finance leaders across the US, UK, India, and Australia, conducted by tax compliance company Avalara, paints a clear picture of this tension. Nearly every executive surveyed said they feel personally responsible for proving that their AI investments are paying off, yet very few say their organization is prioritizing governance over speed. The result is a widening gap between how fast agentic AI is being deployed and how well it is actually being controlled.

Why Finance Leaders Are Adopting AI So Quickly

The case for AI adoption in finance is easy to understand. Agentic tools promise faster processing, fewer manual errors, and the ability to reallocate skilled staff away from repetitive tasks. For finance teams working with tighter margins and leaner headcounts, that kind of efficiency is hard to pass up.

But the research suggests something else is also driving momentum: fear of falling behind. Worldwide spending on AI is projected to climb sharply this year to $2.6 trillion from $1.76 trillion in 2025, and analysts expect that growth to continue for years to come, making the technology too significant to ignore even for functions as tightly regulated as finance. When competitors and industry peers are already deploying agents, standing still starts to feel like the riskier option, regardless of whether the underlying infrastructure is ready.

The Pressure to Adopt AI Is Reshaping CFO Strategy

For many CFOs, the pressure to adopt AI isn't abstract. It shows up in performance reviews, board conversations, and budget justifications. In the survey, the vast majority of respondents said they feel personal career pressure to show that their AI agent investments are delivering against business goals. Half described that pressure as significant, though the intensity varies by region, with US-based leaders reporting considerably more pressure than their counterparts in Australia.

Career Stakes Behind Every AI for CFOs Decision

That pressure hasn't necessarily translated into results. While most finance leaders report at least some measurable return from their AI initiatives, roughly half describe those returns as limited, and a meaningful share say it's too early to tell or that they've seen no clear benefit at all. This is the uncomfortable reality behind AI for CFOs today. The expectation to demonstrate ROI is nearly universal, but the evidence remains thin in many organizations. Leaders are being asked to move fast and show value while the frameworks needed to measure that value responsibly are still catching up.

AI Challenges for CFOs Where Governance Falls Behind

The most pressing AI challenges for CFOs aren't really about the technology, it’s about oversight. Most organizations surveyed admitted they don't have anyone on the finance team dedicated to understanding how their AI agents actually operate. Instead, teams lean on IT departments or the AI vendor itself to answer questions about how decisions are being made, and a notable share have no one responsible for it at all.

The Speed Mismatch Between AI Action and Oversight

One theme that stood out across the research was the mismatch in timing between how AI agents work and how governance responds. AI agents can execute decisions in milliseconds, while compliance reviews and control processes often take hours or longer. Multiple finance leaders interviewed described feeling like they were constantly playing catch-up, trying to build oversight around a technology that evolves faster than internal policy can be written.

This shows up in the numbers, too. Roughly a third of organizations haven't updated their internal control frameworks in the past year to account for AI agents taking or recommending actions, and nearly half say their incident response plans for AI-related issues are either untested or still being developed.

Common AI Adoption Mistakes Finance Teams Are Making

Several patterns emerged as common AI adoption mistakes among the finance leaders surveyed. Many organizations don't require formal safeguards before a vendor's AI agent is allowed to operate. Fewer than a third require documented human review or escalation thresholds, and only around a quarter require documented audit logs showing how an agent reached a given decision. A significant number of leaders say they rely mainly on a vendor's reputation and existing contract terms rather than any independent technical or compliance verification.

Accountability is murky, too. When agents make significant errors, close to a quarter of respondents said responsibility would be unclear or would sit with no one in particular. Liability arrangements between vendors and customers vary widely from contract to contract, leaving many finance leaders uncertain about where the buck actually stops if something goes wrong.

Building an AI-Ready Finance Function

Despite these gaps, few finance leaders believe the answer is to slow down. The opportunity is simply too large to ignore, and most are instead trying to figure out how to build an AI-ready finance function that can move quickly without sacrificing control.

Preparing Finance Teams for AI Without Losing Control

Preparing finance teams for AI starts with ownership. Rather than treating AI governance as an afterthought, organizations that manage this transition well tend to build clear lines of responsibility across finance, IT, and compliance from the outset, with documented processes that don't disappear when key staff move on. Given how many finance departments currently lack in-house AI expertise, the accountability and transparency of the technology partner an organization chooses matters more than ever.

Balancing AI and Human Expertise

Finance leaders were notably more comfortable letting AI agents assist with tasks like reviewing invoices or flagging discrepancies than they were letting an agent act independently on high-stakes decisions, such as releasing a payment. That instinct points to a broader principle for balancing AI and human expertise. Agentic AI works best when it removes friction from repetitive work, while final judgment on higher-risk decisions stays with people. Defining exactly which actions an agent can take on its own, which require human sign-off, and how unusual outputs get escalated, is a foundational step many organizations still haven't taken.

Why AI-Ready Data Matters More Than the Model

One of the clearer findings is how CFOs are drawing lines around general-purpose large language models. A meaningful share of respondents said these models aren't yet appropriate for decisions or actions within regulated financial workflows, and even those willing to use them for compliance-adjacent tasks insisted that every output be reviewed by a person first. The consistent message is that AI-ready data, meaning verified, domain-specific, and well-maintained information, matters more to reliable performance than the sophistication of the underlying model.

Before expanding any agent's role in a regulated workflow, finance leaders are increasingly asking vendors direct questions about where the data comes from, how it's maintained, and how compliance accuracy is verified.

Why AI Governance Must Come Before AI Adoption

The pressure to adopt AI in finance isn't going away, and for good reason: the potential upside in speed, cost savings, and accuracy is real. But this wave of adoption is also exposing how far governance, expertise, and accountability still have to go before agentic AI can be trusted with the decisions that matter most in a regulated environment. Finance leaders who treat governance as something built into their AI strategy from day one, rather than bolted on afterward, will be far better positioned to capture the benefits of AI without inheriting its risks.