Why Finance Leaders Are Right to Question AI and What’s Next

“The leaders who make the greatest impact don’t wait for certainty. They take responsibility when the outcome is unclear and lead anyway.” – from my piece on how leaders think about risk

AI is being pushed aggressively into finance while finance teams are still expected to operate with almost no margin for error. That pressure is hard to ignore when finance teams are juggling responsibilities that encompass compliance, audit, reporting accuracy, security, and cash flow. 

Everyone is asking finance teams about AI. Every software company suddenly has an AI pitch, and virtually every competitor is announcing a new AI initiative faster than you can blink. Needless to say, finance teams are still trying to separate what is useful from what is hype.

In finance, you are more likely to be criticized for being careless than for being cautious. The culture of caution exists for a reason. The challenge, though, is figuring out where AI actually improves financial operations and where it simply introduces more risk. 

The Skepticism Is Backed by Data

A lot of the skepticism around AI in finance is backed by real data, and the research supports their position more than most people realize.

Pew Research Center ran a survey in 2025 that showed half of Americans now feel more concerned than excited about AI becoming part of their everyday lives, and that number has been climbing steadily since they first started tracking it a few years ago. 

On the enterprise side, MIT’s NANDA initiative looked at hundreds of AI deployments across companies and found that 95 percent of those projects produced no measurable return on investment. And when Wolters Kluwer asked over 1,600 senior finance executives what trends they expected to shape their organizations in 2026, fewer than half of them even mentioned AI.

The same concerns kept coming up repeatedly while working on the Trending Now: AI series with Tote + Pears. The marketing around AI is moving much faster than the actual implementation. A lot of finance leaders are skeptical for good reason. Finance teams are supposed to think carefully before introducing more operational risk.

For anyone who wants to see the full picture of where finance teams stand on AI adoption right now, BILL’s 2026 State of AI in Finance report breaks down the data across business sizes, trust levels, and operational outcomes.

Why AI Adoption Stalls in Finance Teams

Most conversations about AI adoption in finance focus on the technology, which tools to buy, which processes to automate, and which vendor has the best demo. But in many organizations, the real issue isn’t the software itself. It’s whether leadership has clearly explained why these tools are being rolled out and how they will actually be used.

As Chris Gerosa, CFO of R&T Deposit Solutions, shared with me, “One of the biggest mistakes leaders make is treating AI skepticism as resistance to be managed, rather than insight to be understood. Doing that dismisses the experience and judgment of the employees that organizations rely on to deliver results.”

The data confirms a stark reality across industries:

• A global study from Great Place To Work found that 85 percent of workers have access to AI, but only 44 percent feel excited about using it or trust their employer to handle it responsibly
• A FranklinCovey survey found that 80 percent of individual contributors describe their manager’s AI leadership as completely hands-off
• Great Place To Work CEO Michael C. Bush framed it clearly: if adoption is slow, it is not an AI problem; it is a leadership problem

In finance, the margin for error is often smaller, and the consequences are often larger. A poorly implemented tool in marketing might waste budget, but a poorly implemented tool in financial operations can lead to compliance violations, failed audits, and reporting errors that come with profound legal consequences.

A Leadership Framework for Getting AI Right in Finance

A few patterns show up consistently among finance teams getting real results from AI, and it connects to what I have written about when exploring how to lead digital transformation effectively. Here are a few principles that matter for finance leaders evaluating AI right now, whether you are a controller, VP of Finance, business owner, or CFO making decisions about AI adoption.

1. Start With the Problem, Not the Tool

Many teams start with the software instead of the operational problem they are trying to solve. The leaders who get this right identify a specific operational bottleneck first, whether that is cash flow visibility, invoice backlogs, or reconciliation delays, and then evaluate whether AI actually solves it. According to BILL’s research, almost two-thirds of businesses cannot access their current cash position on demand, and that is the kind of concrete problem worth solving before you ever open a vendor demo.

Will Taubenheim, founder of Lost Frame Ventures, sees the same pattern from a technical implementation standpoint. He told me, “Do not start by looking at what AI can generate. Start by looking at where your team spends the most manual, repetitive time. Usually, this is in manual data entry, document ingestion, and searching for information across disjointed folders.”

That’s the right starting point for finance leaders. The goal isn’t to bring AI into the organization as broadly as possible. It’s to identify the work that’s repetitive, high-friction, and low-judgment, then use AI to remove that burden without creating new risk. This works because it adds an outside expert voice, makes the principle more specific, and keeps the article focused on leadership judgment rather than turning it into a technical AI piece. The current article already makes this argument in the framework section, so the quote fits naturally there.

2. Build Trust Before You Build Workflows

Before rolling out any tool, address employee concerns directly. A joint study from Duke University and the Federal Reserve Banks found that AI had a negligible impact on actual employee headcounts through 2025 and 2026, and that kind of data matters when employees are worried about what AI adoption means for their jobs. Great Place To Work’s research shows that employees with no AI training still feel enthusiastic about it when they trust their leaders to get them trained at the right time. As Marna Ricker, EY Global Vice Chair – Tax, told me, “Never before has there been a more important window for reskilling and upskilling. Expectations, roles, and personas are changing quickly, and continuous learning is going to be essential.” People are far more likely to use these tools when leadership is clear about why they are being introduced. 

3. Measure Outcomes, Not Usage

Too many companies measure AI success by usage instead of outcomes. The real question is whether the technology is saving time, reducing errors, or improving decision-making. BILL’s 2026 State of AI in Finance report found that finance teams using AI saved an average of 21 hours per week, and 75 percent of leaders reported measurably fewer errors. Those are the metrics that matter.

4. Choose Vendors Who Earn Trust With Data

Not all AI is built on the same foundation. Finance leaders should evaluate vendors based on where the data comes from, how it is secured, and how transparent the methodology is. AI trained on proprietary financial data is very different from models trained on scraped internet content. BILL’s platform draws from data across more than 8.3 million network members, which is the kind of specificity that produces reliable outputs for financial operations.

Step	Key Question	What to Look For
Start with the problem	What bottleneck costs us the most time or risk?	Cash flow gaps, reconciliation delays, invoice backlogs
Build trust first	Does my team understand why we are doing this?	Open conversations about job impact, training timelines
Measure outcomes	What changed because of the tool?	Hours saved, error reduction, faster close cycles
Choose vendors carefully	Where does the AI’s data come from?	Proprietary financial data vs. generic internet scraping

That feedback captures exactly what the conversation around AI in finance leadership requires right now: a framework that connects real leadership experience to the specific pressures finance professionals are facing, not another product pitch dressed up as thought leadership.

What the Best Companies Are Doing Differently

BILL’s 2026 State of AI in Finance report shows that 86 percent of finance leaders have moved forward with AI adoption in finance in some form. So the question is no longer whether to adopt; it is why some teams see real impact while others stay stuck in pilot mode.

The companies seeing the strongest results usually communicate implementation more clearly and give employees more support during rollout. Great Place To Work found that 81 percent of employees at the 100 Best Companies feel psychologically safe at work, compared to just 56 percent at typical organizations. That safety is what makes people willing to experiment, ask questions, and actually integrate new tools into how they work every day.

The strongest companies are not using AI just to cut costs. Many are using AI to improve decision-making and efficiency instead of viewing it primarily as a cost-cutting tool. The companies getting the most out of AI focus as much on implementation and communication as they do on the software itself.

Frequently Asked Questions

Why are so many finance leaders still skeptical about AI?

The skepticism comes from a combination of real data and professional instinct. Great Place To Work’s research shows that only 44 percent of workers trust their employer to use AI responsibly, and MIT found that 95 percent of enterprise AI pilots fail to deliver measurable returns. Finance teams face larger consequences when errors happen, so they are naturally going to approach AI more carefully than most departments.

How can finance leaders build an effective AI strategy without overspending?

If you are a finance leader looking to get AI right without burning through the budget, pick one process that eats up the most time for your team, whether that is accounts payable, receivables, or monthly reconciliation, and start there. Track what actually improves in hours saved and errors caught, not how many people opened the tool. BILL’s 2026 report found that finance teams applying AI to specific bottlenecks saved around 21 hours per week, which is a solid benchmark before expanding further.

What is the biggest mistake finance leaders make when adopting AI?

Deploying tools without first addressing the trust and fear that people on the team are carrying. Research shows that 80 percent of employees say their manager takes a completely hands-off approach to AI leadership, and that disconnect between rollout and support is where most adoption efforts fall apart. The technology usually works fine, and the leadership around it is what determines whether people will use it.

Will AI replace finance jobs?

A joint study from Duke University and the Federal Reserve Banks found that AI has had a negligible impact on employee headcounts through 2025 and 2026, which is consistent with what I have heard from leaders across industries. AI in finance leadership is changing what finance professionals spend their time on, shifting the focus from transactional processing toward strategic analysis, forecasting, and advisory work. For many finance professionals, that shift means spending less time on manual processing and more time on strategic work. 

Disclosure: This article includes insights from research provided by BILL, a company I’m partnering with.