Seven Predictions for Banking, Insurance, and Financial Markets Finance Teams in 2027

My career has afforded me the opportunity to view the office of finance function from two different perspectives. First in roles as a CFO and FP&A leader responsible for forecasts and explaining results and in my current role working with finance teams that are trying to improve their planning environments that have often evolved over many years. That experience shapes how I think about where FP&A is headed.

As 2027 approaches, finance departments have a lot on their plate.  Rates are still hard to predict. Management wants answers faster. Reporting and regulatory requirements keep expanding. And AI is showing up everywhere, in forecasting, reporting, analytics, decision support, pretty much all of it. 

The technology will keep getting better. I’m not convinced, though, that picking the right technology will be the hardest FP&A problem in 2027.

For most CFOs, the harder question is whether they can trust the environment producing the numbers. Can we explain the forecast? Can we trace a number back to its assumptions and source data? Can we quickly change those assumptions? Do we genuinely understand the logic that’s built inside the model?  And as AI becomes part of the process (which it will), do we know where automation ends and real financial judgment begins? 

Those questions are going to matter a lot more than they do today. The reality is many teams are already struggling with them and just haven’t come to terms with it.

Here are seven developments I think will shape FP&A in 2027.

1. Agentic AI gets adopted quickly, yet true autonomy is not in the cards

You will see AI woven into the fabric of forecasting, variance analysis and management reporting with increasing frequency. I do not foresee finance leaders ceding consequential decisions to an autonomous agent so easily.

While a large majority of finance executives are at ease with some form of agentic AI, very few are ready to let it call the shots on important matters.  Expect every serious planning tool to include audit trails and human approval steps.

2. Scenario planning will take precedence over the single-point forecast

There was a time when finance organizations devoted inordinate amounts of effort to nailing down a perfect forecast. By 2027, the more pertinent question will be: what if our assumptions are wrong?  No single forecast can capture that uncertainty.

The better FP&A groups will keep a base case on hand but also have upside and downside scenarios with drivers that are easy for management to grasp and adjust. Boards will want to know about forecast accuracy, to be sure, but they will be just as concerned with whether management had a handle on the range of possible outcomes. It is less about being right on one number and more about understanding the business well enough to anticipate changes.

3. When planning models are left undocumented, it becomes a risk

Does this sound familiar?  There is a model in the planning platform or a spreadsheet that has been around for years. It works and everyone uses it, but the logic is not well documented and only a couple of people really know how the pieces fit.

Where that might have been seen as nuisance in the past, CFOs are starting to treat it as a governance problem. Could you reproduce the forecast if a key person walked out the door tomorrow? Can you trace a result back to the source data and assumptions?

These questions create operational risk. And as AI is layered onto existing environments, they matter even more; automating something opaque does not improve governance, it just makes it happen faster. Documentation, ownership and change control need to be part of the financial control environment, not merely optional housekeeping.

“If the person who built it left tomorrow, could anyone else run it?”

Rusty Pennington · Co-Founder & CFO, Grandview

4. The scope of the CFO will grow to encompass AI governance

The CFO’s role has been expanding for quite for some time and AI is only going to accelerate that trend.  

CFOs are already heavily involved in the enterprise’s technology spending and AI investment decisions.  It is logical. Finance is in the best position question an AI project: What problem are we solving? What is the actual cost once you factor in support, data and infrastructure? Where is the expected financial benefit? What is a scalable tool vs an expensive exercise?

I would expect more CFOs to put formal disciplines in place for managing the ROI of AI in 2027. The objective is not to be cheap with AI but to have the governance to invest where there is measurable value.

5. The line between planning and external reporting is blurring

For the most part, finance has long regarded these as two separate processes, even if related. That is changing as reporting requirements are pulled upstream into the planning function. It is harder to make the distinction these days.

Consider the new accounting disclosure rules. Many public business entities will face expanded expense disaggregation requirements when they put together their 2027 annual reports, which is just one instance of the growing call for more granular financial data. 

The trend is unmistakable: there is an expectation for organizations to be able to understand and account for their results in fine detail.

A good approach is to build your planning models from the ground up with the right dimensions, classifications and supporting detail already in evidence. I would expect top finance teams to be asking themselves what information they will need down the line to justify a given number, and to structure their planning architecture around that.

6. Excel is here to stay, but no longer as the system of record

Finance people are never going to stop using Excel. Nor should they. It remains one of the best tools available for analysis, modeling and presenting plans or results. 

The question is where the governed logic is to be found. You should not have critical business rules, security and knowledge locked in a disconnected spreadsheet that is hard to govern or replicate.  I would expect the norm to become a straightforward approach: let Excel be the interface for the user, but ensure that the calculation logic, version control, workflow and auditability are housed in the enterprise planning environment. That is an important separation.

7. The maturity of FP&A can be judged by self-service analytics

The strongest FP&A teams give business units real self-sufficiency in accessing reports and data. We always evaluate headcount, revenue and spending but one measure of success should be how many people outside of the finance group can get trustworthy data and answers without contacting the finance group.  

Self-service is not about letting operating leaders run wild with the data or build their own models. It is about having a controlled setting where an authorized person can get a straight answer to a business question without filing a ticket with finance or having to argue over which version of a spreadsheet is correct.