CFOs remodel the finance function for an AI-driven world

From ROI goals to big bets

As AI adoption reaches a critical mass, Chief Financial Officers (CFOs) are striving to measure AI’s true value and build convincing business cases.

Some CFOs are finding success by taking calculated risks and forcing change, while other leaders are struggling to justify major investments to their more conservative boards. Many more are stuck in what CFOs refer to as the messy middle.

The CFO for a global media and information services firm described their “necessity is the mother of invention” approach to AI.

“We set a hard deadline. By this date, we are going to change the way work gets done. Our team built a flux analysis agent in three weeks. We went from having 40 people manually grinding through variance commentary every month. That’s down to five people reviewing and refining what the agent generates. It happened because of the deadline, and we gave them the runway to build it.”

At the other end of the spectrum are finance leaders facing pushbacks internally. A CFO noted the challenge of getting the green light for projects without a clear, multi-year payback. Other leaders rely on softer proxies to get projects approved.

For example, after years of failing to get approval, a CFO in the aerospace and defense sector scored with a business case around hard efficiency savings from automating manual workflows. It looked good on a spreadsheet.

Most companies fall within what finance leaders classify as the messy middle. In the automotive industry, a prime example of this evolution is a company’s AI journey that began three years ago with a goal of making employees’ lives better. That led to significant efficiency gains in year 1 with 100,000 hours saved to year 2, 200,000 hours.

This year, the company wanted more results. To achieve this, they identified 10 big bets. These are specific, high-impact AI initiatives focused on tangible financial returns. For example, using AI and large language models to scrape pricing data and optimize the parts business saved $100 million this year alone.

All told, the 10 big bets are projected to deliver $500 million in EBIT this year. It demonstrated a transition from a messy AI strategy to a results-based one.

“Here are 100 great ideas. We pursued three. The other 97 ended up on the cutting room floor.”

CFO, insurance sector

AI is redefining a successful career in finance

For many finance leaders, AI has automated entry-level tasks, leaving many CFOs wondering how the next generation will ever learn the business. Meanwhile, experienced performers have witnessed their ways of working competing with AI-native new hires who think and work differently. There is a palpable tension between generations. 

Case in point: an airline CFO shared a story from his company. An intern in a meeting casually corrected senior team members on what AI models to use for a specific task. The intern said, “Why are you starting with CoPilot? Claude would be much better for this.” 

For mid-level managers who cut their teeth on Excel, there can be a sense of numbered days given the new wave of AI-native talent and the company’s embrace of AI.

In response, CFOs are proactively redesigning their talent playbooks. An automaker CFO rolled out a capability framework in finance around three themes: thinking enterprise-wide, turning data into action, and driving results with lean thinking. This also includes mandatory AI fluency training that is now tied to career progression.

Other companies are experimenting with rapid rotations, moving key hires through four different jobs in two years to jumpstart development of broad business knowledge. 

A CFO in the retail sector is also piloting a business rotation model that places new finance hires in business units first. There they learn the business before they are taught specific finance skills.

Erin McShane, managing director, KPMG Human Capital Advisory, sees AI and automation supporting financial planning and analysis (FP&A), spreadsheet modeling, data extraction, and variance commentary. Talent needs are less about a “reporting factory” and more about interpreting the different AI outputs and advising leaders.

Erin also sees the career path for finance evolving. “There will be less emphasis on role-based talent strategies. It’s going to be more about skills and capabilities. It’s almost like earning badges.”

An insurance company CFO believes that functions like controllership, tax, and audit will remain distinct even with AI integration in finance.

In this new world of finance careers, the most prized attributes are strategic inquiry, interpretive storytelling and judgement under ambiguity. In other words, the unique human ability to make critical trade-offs when AI surfaces patterns. 

“There’s a lot of very defined silos in finance that might be much less defined in the future.”

Erin McShane, Managing Director, KPMG Human Capital Practice

From step changes to radical moves

Facing cost pressures and tantalized by AI’s capabilities, CFOs are reshaping finance, from incremental improvements to rethinking their standard operating procedures.

A much-shared desire among finance leaders is to end the practice of shadow finance units that have proliferated across business units. The CFO of a financial services firm captured the frustration stating, “Amen to the shadow finance function. I’m trying to smoke them out. We get stuck reconciling it; it’s my problem.”

The consensus among CFOs is that now may be the time to end duplicated work and consolidate into a single source of truth or package it as enterprise intelligence. It would be owned by finance with the mandate to eliminate waste.

The embrace of thinking bigger and the day-to-day drive to be more efficient is compelling CFOs to question how much their companies spend on technology. For example, there is a growing aversion to expensive, long-term software contracts.

“Every time someone comes to me wanting to fund software,” shared an automotive CFO, “I just ask, ‘are you sure we can’t do this with AI?’”

This view was set into motion by how fast AI is progressing in its capabilities. The long-held belief that data must be optimized for AI is being called into question. The CFO shared their experience.

“We’re able to do some remarkable things that we could not do even 9 or 12 months ago. We’re on a pathway to clean up all our data. Some of our data is in COBOL. There is less of a need to clean up the data. We can ingest COBOL code right into AI as is and convert it to Python.”

Stories like this are signs that CFOs are willing to make bold structural changes to create a leaner, more agile, technologically empowered finance function.

A CFO in the specialty materials sector, tasked with a major cost-cutting initiative, described his mandate as “untangling the spaghetti bowl of processes.” They explained. “We’re looking at end-to-end processes like procure-to-pay and quote-to-cash and questioning, ‘Do we have the right operating model?’”

Not long ago, such radical changes would have been unheard of in finance. AI has opened the floodgates of change.

“"Every time somebody wants to spend more on software, I just ask, 'Are you sure we can't do this with AI?'"

— CFO, automotive sector