Artificial intelligence is quickly becoming one of the most consequential tools available across every aspect of business, and the finance function is no exception. Yet many CFOs report that the path from promise to positive outcomes remains anything but straightforward.
While use of AI in areas like automating routine workflows, improving forecasting and surfacing anomalies is well under way at some organizations, others are still wrestling with foundational challenges such as data readiness, use-case prioritization and internal expertise. The result is a landscape where both degree of implementation and operational impact vary widely.
“People are all over the map for sure on how they’re adopting this,” says Dan Owens, CFO of Maxio, who, despite the hurdles, sees leaning into the AI learning curve as essential. “If people aren’t embracing it, they’ll be the ones that fall behind—and the ones embracing it are going to get more efficient.”
Mitch Ginsberg, founder of CommLoan, seconds that sense of urgency. “Either you’re going to ride the technology wave or you’re going to get swept away by it,” he says. “AI is already having immediate, enormous effects on our economy, and it’s just beginning. It’s a huge mistake for anybody in any industry to ignore it.”
As expectations around AI continue to rise, the question is not whether the technology can deliver value to the finance function but how CFOs can best approach laying the necessary groundwork to support it, choosing the right use cases and moving from experimentation to successful outcomes. The finance chiefs and AI experts we canvassed outlined six steps:
Start with the problem, not the solution
The race toward adoption can lead CFOs to rush into exploring AI capabilities before identifying the business needs where it can deliver value. Veterans of the AI adoption process stress that the starting point should always be the problems, processes and opportunities, never the technology.
“Before you even go down the AI road, make sure you’ve done a robust level of strategic planning to find out the areas where you can definitely get more value, more productivity and where AI can help you with that,” says Rob Goldstein, a partner at CPAs by Choice who guided Arnold Packaging through automation initiatives in his previous role as its CFO. “Otherwise it’s like saying, we’re going to revolutionize our company by putting a computer on everyone’s desk. Well, what good is that computer? What are you trying to solve, and how is the computer going to help you do it?”
“The mistake many companies make when thinking about AI in the finance function is starting with the technology instead of the decision-making,” agrees Maria Pearman, who serves as a contract CFO to businesses in the food and beverage sector. “When I evaluate AI and automation for finance, I start with a simple question: What decisions do we wish we could make faster or with more confidence? That usually leads us to three areas where the ROI is immediate—close processes, forecasting and data integration across systems.”
How AI Can Help You Become A Growth-Oriented CFO
To move from counting value to creating it, CFOs must shift from manual record-keeping to AI-native systems that can predict, reason and act.
Advisors echo that approach. Ric Opal, global digital leader at the professional services firm BDO, urges CFOs to “take technology off the table” when embarking on their AI journey, instead focusing first on friction points in their business. Where are there bottlenecks, growth constraints or operational inefficiencies? From there, the goal is to identify “one or two, not 10” problems AI can help address.
These projects should be quick, iterative wins, not multi-year transformations, notes Opal. “We’re looking for the quick strike we can get done. We anticipate jointly that it will not be perfect from the first day… and then we collaborate to make sure we get over the line and get to a good outcome.” Delivering that early first win then sets the organization on a path from which to build, iterating as capabilities and confidence grow.
Build a clean, connected data foundation
AI systems rely on access to large volumes of accurate data in a digestible form, yet many finance organizations still operate across fragmented systems with inconsistent data definitions. “The output is only as good as the input,” says Chris Gerosa, CFO at R&T Deposit Solutions. “If your data is not clean and complete, you can slice and dice it in an automated way as much as you want, but it won’t give you an accurate outcome. So making sure your data is clean is a really important milestone to overcome before you start rolling out AI.”
For many organizations, this means investing time upfront in data quality, standardization, integration and governance—work that may delay AI deployment but ultimately determines its success. Before launching an AI initiative, for example, CFOs should ensure that standardized definitions for metrics are in place across business units and that integrated systems allow for automated data flows, reducing the need for manual data extraction and manipulation.
At the sales enablement software company Seismic, that journey was a multi-year endeavor, says CFO Evan Goldstein. “We’ve spent the last four or five years really looking at our data structures and making them consistent, aligning them where there were deviations, cleaning those up and automating those processes,” he explains. The work gave his team confidence in their data and, over time, the ability to move more aggressively with AI. “If I knew there were anomalies, if I knew teams were downloading data and manually fixing it in spreadsheets, I wouldn’t feel great. But we don’t do that anymore.”
For those that invest early, standardization and data quality creates a compounding advantage, improving reporting and decision-making today and enabling faster, more confident adoption of AI and advanced analytics down the line. “In many ways we’re back to the early stages of ERP implementations, when everybody was talking about ‘garbage in, garbage out,’” says Scott Rottman, president of the CFO advisory arm of the professional services firm RGP.
“Organizations need to have their house in order before they can even start talking about AI. If you jump into AI and try to figure out AI use cases when you don’t have good data governance and data quality around organizational information, your AI mandate will fail. That stuff has to be cleaned up first before you start.”
Put the right people on the right path
CFOs also stress the importance of getting the right talent in place and aligned on moving AI initiatives from concept through execution. Often, that requires making a deliberate investment in upskilling existing finance staff, hiring new talent with data or analytics expertise or carving out dedicated roles focused on automation and AI.
While external advisors can play a valuable role, relying too heavily on outside vendors without considering internal accountability can lead to misalignment and missed opportunities, adds Gerosa, who designated a senior team member as the internal leader responsible for overseeing R&T’s automation initiative with NetSuite AI. “That was an important decision we got right because you need somebody in-house who has building it right as their full-time job. Your people know the business, and they’ll be instrumental in making sure you build it appropriately.”
AI initiatives can also falter if employees are hesitant to adopt the tools due to uncertainty around how it will change roles and expectations. “In finance, people worry—is this going to take my job?” says Denise Graziano, CEO of Graziano Associates, who says that without clear communication, that concern can lead to reluctance to engage with new systems, jeopardizing success.
Graziano advises addressing employee concerns head-on through transparent, consistent communication about why AI is being introduced, what problems it is meant to solve and how roles are likely to evolve. “Leadership needs to explain how it’s going to change people’s jobs—because it will—and how they’re going to be upskilled or reskilled so they can stay relevant,” she says. “The people who get it right are the ones who are honest about it, which may mean saying, ‘I don’t know exactly what any of our roles are going to look like a year from now, five years from now, but this is our plan and here’s why.’ When employees are brought in on the decision, they are more invested in using it, and its success becomes a team effort.”
Get your governance framework in place
Companies that advance AI initiatives without putting policies and procedures in place around its use risk running into trouble down the line. The lack of guardrails introduces a wide range of risks, such as employees publicizing intellectual property while experimenting with AI or running afoul of regulatory bodies. And once the absence of controls introduces vulnerabilities, backfilling with financial controls can be hard, says Gerosa. “Having a consistent, strong set of policies, procedures and a risk framework across disciplines in place is critical because trying to add it later or do it in parallel when you do the system will be really hard.”
To tackle the governance challenge, Seismic set up a cross-functional AI Council populated with stakeholders from product, legal and compliance, and information security, along with a designated leader to guide AI adoption. Early on, the group focused on surfacing and vetting product-oriented AI initiatives aimed at meeting customer needs, but over time the role broadened, says Goldstein. “Now we’ve said, ‘We need a governance framework internally in how people use AI,’ and they’re going to be tasked to derive that and present that to the leadership of the company so we can debate that and move forward on it.”
Such a model creates shared ownership, helping to break down silos and ensure that decisions about AI are informed by multiple perspectives, not just finance or IT. It can also help a company develop a clear framework for acceptable use, data access, model validation and escalation protocols, as well as practices for evaluating potential use cases and vetting vendor partnerships.
Identify workflows where AI can deliver value
Once the foundational elements are in place, finance chiefs can begin exploring areas where AI has the potential to reduce friction, create efficiencies and improve decision-making. Many turn first to financial planning and analysis, where AI can improve the speed and quality of decision-making.
“One you start to ingrain leveraging driver-based analysis, anomaly detection, predictive analytics—all of these capabilities—into your process, not only are you driving efficiency, but you’re in a position as a finance leader to close the books quicker and get to that forecasting piece earlier to drive a more accurate forecast,” says Mike Shuker, global head of solution consulting at Wolters Kluwer. “That will translate to your bottom line and also enable you to place your bets quicker in terms of where you want to deploy capital and what key initiatives you’re going to drive, whether that’s in supply chain or operations. And if you get any bad news early, you’ve got the agility to move. It gives you a lot more options.”
Several CFOs report initiatives that are helping their teams streamline time-sensitive processes. At Seismic, the finance team leans on AI to smooth deal approval and pricing workflows. The company developed an AI agent that pulls relevant data from previous deals—deal structure, approval records and pricing decisions—to assess proposed deals and flag those unlikely to make it through the approval process.
Final approval still rests with finance leadership, says Seismic’s Goldstein, but the output allows the sales team to adjust expectations earlier in the process and avoid bottlenecks during crunch periods. “Sales reps push us really hard at the end of the month, so it’s helpful if we can tell them a little faster, ‘Hey, this is not likely to get through.’ That way we’re not in paralysis with our reps pushing for something that’s not going to happen.”
AI can also fill in the gap for companies that lack a dedicated FP&A resource. “Right now, our vice president of finance and I tag-team on that function,” says Owens, who says Maxio developed an AI-driven analytics tool that users can query in plain language, effectively turning the reporting dashboard into a conversational interface. “I’ve been so impressed with the quality that it’s made me question whether we need to hire a dedicated resource. Once you train it, the system can produce outputs comparable to what you’d get from a mid-level FP&A professional.”
Maxio is also using AI to surface the most likely drivers of customer dissatisfaction or attrition with an eye toward improving retention. By analyzing data across the full customer lifecycle—“from Zendesk [support] tickets to Salesforce [CRM] interactions to Gong [conversational intelligence] calls, AI can identify the top five attribution reasons that a customer may not be happy,” says Owens. “Most companies collect all this but don’t have a way to aggregate it and use it proactively to correct a bad experience. With AI you can do that and you do it in a less biased way than asking, say, your salespeople, who will tell you that it’s a product problem, or your product people who will say it was implementation, because no one wants to blame their own function.”
Start, and then expand on, experimentation
For many finance leaders, an AI roadmap is coming into focus: Start with a defined problem. Invest in the underlying infrastructure—data, processes and people—to support it. Focus on targeted use cases where AI can accelerate decision-making or reduce friction. Then build from there, iterating as capabilities and confidence grow.
Success with AI, CFOs say, will come less from sweeping transformation efforts than disciplined execution grounded in clear business priorities, strong data foundations and thoughtful governance. Finally, AI adoption is not a one-and-done initiative, but an ongoing capability that will evolve, requiring the CFO role to morph along with it, moving from overseeing financial reporting to orchestrating how data, technology and talent come together to drive performance.
Ultimately, in finance as elsewhere, the AI winners will not necessarily be those who move first, but those who move with purpose.
