Beyond the annual budget: Rolling forecasts in the age of AI

How many of us have experienced building a budget, only to realize—no sooner than it’s approved—that all the numbers have changed, and it’s time to reforecast? Budgeting is as much about the planning process as it is about the numbers. However, if we never change the budget and it becomes unachievable shortly after it was created, that can become demoralizing for everyone involved.

The truth is, we need numbers and targets that reflect reality, and that’s why rolling forecasts are an effective way to keep plans aligned with changing business conditions and move beyond the limitations of a static annual budget.

“The budget evolved from a management tool into an obstacle to management.”

Frank Carlucci, Former U.S. Defense Secretary

As rolling forecasts are becoming increasingly important, in this article we’ll start by defining what a rolling forecast is and then discuss three key areas to focus on to successfully implement a rolling forecast process at your company:

1. Cadence
2. Drivers
3. Accountability

Within each of these areas, AI is a magnifier and, when used correctly, can help Finance teams detect change sooner, identify key business drivers, and improve forecast quality. However, AI is most effective when paired with strong processes and knowledgeable Finance professionals who provide context and judgment.

1. Cadence: Forecast when the business needs it

“The goal of forecasting is not to predict the future but to tell you what you need to know to take meaningful action in the present.”

Paul Saffo, Futurist

Historically, many companies have reforecasted their businesses whenever the calendar dictated. For example, a reforecast was done once a quarter or every month, regardless of how accurate the existing forecast was or whether the business needed new numbers. This is what we call a calendar-driven forecast, and in today’s environment, it’s insufficient. Having a regular cadence for forecasting is a starting place, but from there, we need to move toward event-driven forecasting.

The reality is that when uncertainty increases, forecasting cycles should adapt to business conditions rather than fixed schedules. The prime example of this is COVID, when new forecasts were produced weekly and, in some cases, every couple of days. Building robust processes that allow you to forecast when events demand it is important in today’s environment. This requires using a planning tool or the extreme discipline of the business and Finance function to do this in Excel.

The principle behind event-driven forecasting is being able to meet the business’s needs. If you normally forecast every month but nothing has materially changed, skip that month and don’t forecast. If other priorities have come up that are more important than the forecast, focus on them first. For example, I used to have months when the forecast and paying our Sales team conflicted; I always chose to pay the Sales team because that was more important than an arbitrary forecast deadline.

However, if the business needs a new forecast mid-month due to underlying changes, you must be ready to reforecast. The trigger events can be external, such as a rise in commodity prices or a competitor’s move, or internal, such as a new project launch or the loss of a major customer. A good rule of thumb is to be ready to reforecast when the event will materially change the business forecast. What’s material will vary by business, but a change of more than 5% is often a good place to start.

The core principle of cadence-based forecasting is to meet the needs of the business, rather than relying on an arbitrary calendar to decide when to reforecast. As you decide when it makes sense to reforecast, remember it starts with listening to your business leaders and goes beyond that.

Building in alerts using AI and technology to let you know when things have materially changed can be very helpful. Sometimes waiting for the business to signal the change can be too late.

For example, modern AI tools can continuously monitor internal and external data sources, identify emerging trends, detect anomalies, and highlight forecast assumptions that may no longer be valid. This allows Finance teams to move from reactive reforecasting to proactive decision-making. As you move to an event-driven forecast, it becomes more important than ever to have an efficient forecasting process that focuses on the key drivers of the business.

2. Drivers: Focus on the drivers that move the business

Many different methods exist for forecasting a business, but one thing all good forecasts have in common is that they understand the operational drivers that move the business forward. Building a forecast focused on accounting line items and simple averages makes it nearly impossible to work closely with the business to achieve the targets it committed to. To do this, you need to translate the forecast and key operational drivers into performance metrics that can be tracked and reported.

The key to making driver-based forecasting work is as follows:

1. Tying each driver back to the operations of the business
2. Selecting drivers that the business can easily influence and track
3. Focusing only on key drivers (3-5) that move the business

Let’s look at a simple example of two forecasts, both designed to achieve the same revenue target but using very different paths. In this example, we have an e-commerce business that sells outdoor gear.

In this first chart, we see that the business has used a simple growth assumption for units sold and an inflation assumption for price. This results in a plan number of $12.9M, but it doesn’t provide details to hold the business accountable. We have no idea how much website traffic or what conversion rate we need to achieve to reach these targets. This simple plan, built on basic assumptions, doesn’t help the business understand how we intend to achieve it. Contrast this simple forecast with the example below, which is focused on forecasting using key drivers.

In this example, we start the forecast by examining key metrics, including organic website traffic and the baseline conversion rate. We then make our key plan assumptions, including inflation, website growth, conversion rate improvement, and the traffic split by category. Building the plan with key drivers makes it much easier to hold the business accountable. Say, for example, in February, we miss the target. We can review the key assumptions and learn that the conversion rate was only 2.7%, whereas we had assumed 2.9%. This not only tells us what we missed but also allows us to have conversations with the business to determine the root cause of the miss and what to do about it.

When it comes to deciding what the key drivers should be for our forecasting models, we can use AI to help us understand what key operational drivers might be for our business. For example, if we’re new to the business we’re supporting, we could use Generative AI to research the industry and summarize the key operational drivers for businesses in our industry. We could also use it to analyze our competitors’ recent 10-K and 10-Q filings and to research online articles about our industry. In addition to AI, it would be smart to speak with the subject matter experts (SMEs) within the company and use tools such as value driver trees to map out the key operational drivers.

Using the driver-based approach to forecasting makes it much easier to update rolling forecasts, as we only need to work with the business to update the key assumptions that drive the forecast, not every line item. A driver-based forecast also sets us up nicely for our next key to rolling forecasts, which is forecast accountability.

3. Accountability: Turn forecasts into a shared responsibility

“Accountability is the glue that ties commitment to results.”

– Bob Proctor, Author

All too often, after the budget or forecast is closed, Finance hears some variance of the following in a meeting: “That’s not my number. I have no idea how Finance came up with it.” More often than not, when something like this happens, it’s a sign that the business is not being held accountable to the forecasted numbers they signed up for.

The problem is rarely about control; more often, it is about a lack of process, visibility, and trust. If FP&A is to help the business achieve its financial targets, it must create a culture of shared accountability. All too often, business leaders look at the budget/forecast as Finance’s number, not a shared target that everyone helps achieve.

When creating the budget, FP&A teams should work closely with the business to agree on the key drivers for major revenue lines and departments. For example, for the implementation team, the key operational driver might be the time it takes to complete an installation of a new software. If this is the case, make sure you work closely with the implementation team and align on the estimated installation time, the key variables that could affect the targeted time, and the major risks and opportunities.

One simple way to improve accountability is to assign an owner to each major forecast driver. Sales may own conversion rates, Marketing may own website traffic, and Operations may own fulfillment efficiency. When ownership is clear, discussions move beyond explaining variances and toward improving performance.

As you work to create an environment of shared accountability, make sure, as a Finance professional, you focus on being collaborative and supportive of your business partners. One of the fastest ways to kill trust and accountability is by viewing everything through the lens of Finance vs. the business, instead of Finance and the business. For example, when a target is missed, never say you missed your target; say we missed our target. This shows that you’re part of the team and share in the accountability.

Conclusion: Building a future-ready forecast process

In today’s world, FP&A needs to build a forecasting process that’s centered around the needs of the business. This requires being able to reforecast when the business needs a new forecast, not when the calendar says it’s time.

That new forecast needs to be built around the company’s operational drivers and what moves them. Accountability for those key drivers must reside with the business. Finance is responsible for ensuring the forecast is realistic and working with the business to remove roadblocks to achieving the financial objectives, but the business is responsible for ensuring the drivers used in the forecast are ones it can influence and control.

Technology, including AI, can improve forecast speed, accuracy, and visibility, but it doesn’t replace the fundamentals of strong planning and business partnership. The best FP&A teams treat technology as an enabler that helps them better support the business while keeping a human in the loop. AI serves as a magnifier throughout the entire process. It can alert you when something has materially changed, run the first pass on your variances, and make the numbers visible to the business. However, it takes a skilled FP&A professional to decide what to do with that information. Remember, our goal is to enable the business to achieve its objectives, not just to report the P&L.

If you’d like to see how Jedox enables rolling forecasts in the age of AI, you can schedule your demo here.