A few months ago, I was talking with a finance leader about AI. Like many conversations these days, we started by discussing copilots, automation, and the growing list of tools promising to transform finance. At one point, he laughed and said, 

"That's great, but my team still spends half its time pulling data together every month." 

That comment stayed with me because it reflects a reality that many finance organisations continue to face. While the conversation has moved toward AI, the day-to-day experience for many teams looks remarkably similar to what it looked like a few years ago.

Data still sits across multiple systems and often tells different stories depending on where you pull it from. Analysts spend hours exporting files, reconciling discrepancies, validating numbers, and rebuilding analyses that already existed last quarter. Talented professionals who were hired for their judgment and business acumen frequently find themselves spending more time stitching spreadsheets together than helping leaders make decisions. The problem is not a lack of technology. Most finance organisations have access to more tools than ever before. The real question is whether the underlying operating model has evolved at the same pace. In many cases, I do not believe it has.

For years, organisations have treated AI as something to bolt onto existing processes or as an isolated experiment running on the edges of the business. What often gets overlooked is a more fundamental question: does the way finance operates need to change? I believe the answer is yes. The shift that matters most is not from spreadsheets to AI. It is from building reports and one-off automations to building products: durable, governed systems that capture how the finance organisation actually thinks and operates.

The Talent Tax: A Hidden Cost

Most finance organisations carry a hidden cost that rarely appears on a dashboard and is almost never measured directly. Consider a senior FP&A leader whose understanding of the business is trusted across the organisation. Let's call her Jane. She is the person leaders rely on when performance needs to be explained, risks need to be assessed, or difficult decisions need to be made. Her value comes from interpretation, judgment, and experience. She understands what the numbers are saying and, just as importantly, what they are not saying.

Yet a large portion of Jane's week is not spent applying those skills. Instead, she is pulling data from multiple systems, reconciling unexplained gaps, validating reports, and rebuilding analyses that were already completed in previous cycles. Much of the business logic exists only in her spreadsheet, while much of the context exists only in her head. Each reporting cycle effectively starts over.

This is what I call the Talent Tax. It is the gap between the work people are hired to do and the work the operating model actually requires them to perform. The cost is not limited to wasted hours. It also shows up in something I think of as lost Compounding Intelligence. Every month, experienced finance professionals apply judgment, context, and business understanding to solve problems. Yet that expertise rarely becomes part of the organisation's infrastructure. Instead of accumulating over time, it disappears into spreadsheets, presentations, and individual experience, forcing the organisation to relearn the same lessons again and again.

From Jane to Jane 3.0: What the Shift Actually Looks Like

I have seen this transformation happen in practice, and it rarely begins with a new AI tool. It starts with a much simpler question: what does Jane actually spend her time doing, and how much of that work truly requires her expertise?

In one finance organisation, the leadership team conducted a simple time-tracking exercise during a planning cycle. The results were eye-opening. More than half of the senior FP&A team's time was being spent on data gathering, reconciliation, and report preparation rather than analysis and decision support. This finding is consistent with the 2026 FP&A Trends Survey, which shows that 47% of FP&A time is still spent on data validation and collection. Rather than launching an AI initiative, the team focused first on understanding the process. They mapped the data flows behind their most critical monthly report, identified where manual intervention was required, and documented business rules that previously existed only through tribal knowledge.

That work became the foundation for a data product that automated extraction, validation, and reconciliation activities. Within a few reporting cycles, the team reclaimed a meaningful portion of senior capacity. This was Jane 2.0. Instead of spending her time building reports, she was reviewing them, applying business context, and focusing on forward-looking analysis. The work became less about producing information and more about interpreting it.

Jane 3.0 represents the next stage of that evolution. Once data is connected and business logic is embedded into reusable systems, AI can begin applying that logic continuously. It can identify anomalies, generate draft commentary, surface emerging risks, and highlight issues before the monthly close. Jane's role shifts from producing answers to framing questions and guiding decisions. Her expertise does not disappear. Instead, it becomes embedded within systems that amplify its impact across the organisation.

The Economics of Abundance

There is also a larger shift happening in the background. For most of modern history, human reasoning was a scarce resource. Expertise could not easily be replicated, scaled, or distributed. Increasingly, that assumption is changing. As AI capabilities continue to improve, intelligence itself becomes more accessible and more abundant.

If intelligence becomes widely available, the constraint moves elsewhere. The bottleneck is no longer the ability to reason through a problem. The bottleneck becomes structure. Organisations need connected data, documented logic, clear business rules, and well-defined questions. The companies that benefit most from AI will not necessarily be the ones with the most advanced models. They will be the ones that that create an environment where those models can operate effectively.

The Three Pillars of the New Model

The most effective operating model I have seen is built on three interconnected pillars: Data, Product, and AI. In practical terms, this means finance must define who owns business logic, how recurring decisions are governed, and how critical knowledge is embedded into systems rather than retained in individual spreadsheets.

Data provides the foundation. It creates trusted, governed, and automated information flows that allow teams to stop debating numbers and start acting on them. For FP&A leaders, the critical responsibility is to translate business logic, planning assumptions, and decision rules into reusable systems that the organisation can trust and improve over time. Product thinking changes how finance capabilities are built. Instead of creating isolated reports and one-time solutions, organisations create reusable systems that improve over time. AI then sits on top of that foundation, applying reasoning at scale and extending the reach of the team's expertise.

None of these pillars works particularly well in isolation. AI without trusted data creates noise. Data without product thinking creates fragmentation. Product thinking without intelligence limits scale. Together, however, they create a finance function that becomes smarter and more effective with every cycle.

Building Scalable Systems

Spreadsheets remain one of the most powerful tools available to finance teams. I have seen incredibly sophisticated solutions built entirely within Excel, and those capabilities should not be dismissed. The challenge is that spreadsheets were never designed to serve as institutional infrastructure.

When critical logic lives inside a spreadsheet, the capability often lives with the individual who built it. When that logic is embedded into a product, it becomes part of the organisation itself. The knowledge survives personnel changes, scales across teams, and improves over time. Rather than recreating analyses every quarter, organisations build systems that continuously learn and evolve. Insights compound instead of disappearing.

This is where finance begins to move beyond being a reporting function. It starts becoming something closer to become more of a decision platform: infrastructure that helps the business understand, evaluate, and act.

The Path Forward

None of this is ultimately about AI tools. It is about how finance chooses to operate. The transition from isolated analyses to continuous intelligence is not primarily a technology project. It is an operating model decision.

The objective is not to reduce the size of the finance organisation. The objective is to increase its leverage. The most effective finance teams spend less time assembling information and more time helping the business understand what actions to take. They move from reacting to events after they happen to helping shape decisions before they do.

The organisations that make this shift successfully will not necessarily be the ones running the most AI pilots. More often, they will be the ones willing to make the less glamorous investments in connected data, documented expertise, and scalable systems. They will build infrastructure that operates continuously rather than relying on people to rebuild the same work every month.

Finance's next chapter is not about producing better reports. It is about building the infrastructure that decisions run on.

Where FP&A Leaders Can Start Today

This does not need to begin as a large transformation program. In fact, the most successful efforts often start small.

Begin by auditing a single critical process from end to end. Map where the data comes from, where manual intervention occurs, and where business logic exists only in someone's head. The exercise will often reveal more opportunities than expected.

Next, identify a recurring decision that your team makes regularly and document the logic behind it. It might be a forecast adjustment, a risk threshold, or a reforecast trigger. The goal is not to automate judgment. The goal is to make the underlying reasoning visible, repeatable, and scalable.

Finally, measure the Talent Tax. Have your team track how they spend their time during a reporting cycle. Most leaders are surprised by the results. Once those numbers become visible, the case for change often becomes much easier to make.

The future of finance will be determined by who builds the operating model that allows AI, data, and human expertise to work together most effectively. The organisations that succeed will be the ones that capture knowledge, scale judgment, and create systems that get smarter over time.

Source:

1. 2026 FP&A Trends Survey: https://fpa-trends.com/fp-research/2026-fpa-trends-survey-how-ai-testing-fpas-foundations