The pressure to update financial forecasts faster requires finance teams to understand changes in the business sooner. Automation can accelerate the flow of data, but its broader value depends on whether teams use the time gained to investigate shifts in customer demand, challenge assumptions and inform decisions while there’s still time to act.
Benchmarking data from the American Productivity & Quality Center shows that the median organization reports creating 28% of its financial forecast through touchless processing — meaning no human intervention — compared with 20% at the 25th percentile and 30% at the 75th percentile. The responses from 2,142 companies provide a reference point for assessing the percentage of a financial forecast created through touchless processing.
Deciding where to expand automation requires a closer look at what people contribute to the forecast. Retrieving reports and transferring data offer opportunities to reduce manual work, while interpreting a customer’s changing plans or testing an operating assumption may require direct conversations with business leaders. CFOs can examine those activities to assess where technology could take on more work and where human involvement helps make the forecast reliable.
Look for opportunities in data collection
A forecast may draw on financial statements, accounts payable reports, inventory records and projections from across the business. Repeatedly retrieving reports, adjusting filters and transferring figures into a model can leave substantial manual work between otherwise automated systems.
The opportunity depends less on the number of manual steps than on which ones consume time, introduce errors or hold up the process. At APQC, the finance team used Microsoft Power Query to bring data into the forecast, reducing manual inputs and associated corrections. Setting up and validating the process took substantial effort upfront. Once built, it reduced the work required for each forecasting cycle.
AI agents offer another way to handle retrieval. With the necessary access and connections, they can locate reports, apply filters and transfer information. These tasks require clear instructions about sources, reporting periods and destinations, along with validation of the output. Finance expertise remains necessary to identify missing information and handle exceptions as reports or systems change.
CFOs can ask their teams which recurring tasks follow established rules and which require a fresh decision. Automation can also support decisions by speeding up what-if analysis. Finance still needs to determine which assumptions to vary and how to interpret the results.
Keep business context in the forecast
An account manager may learn that a major customer plans to cut spending before that change appears in orders or revenue. Finance needs to understand the likely size and timing of the reduction to decide how to adjust the forecast.
Focus conversations with business leaders on changes that could materially affect the forecast. What explains an unusual expense? Does projected growth require additional staff? Have those costs been included? These questions help finance connect departmental assumptions that might otherwise be considered separately.
Reviewing material assumptions with the managers responsible before incorporating them can also reduce avoidable corrections. A known expense disclosed late may force finance to revisit work already completed, causing preventable delay. Revisions prompted by new information, however, remain an important and necessary part of the process.
Automated data collection still leaves finance dependent on colleagues who know about customer plans, commitments and operating constraints that systems have yet to capture. Retain those conversations where the information could change the forecast, and establish when such information must be communicated to finance each month.
Check whether automation improves the forecast
Validate each new automated data flow against its source. Over subsequent forecasting cycles, track preparation time, recurring errors and forecast accuracy alongside the percentage created through touchless processing. These measures help assess whether the process is becoming faster and more reliable, although changes in accuracy cannot necessarily be attributed to automation.
When a forecast misses, identify the cause before changing the process. An incorrect data feed needs a technical fix. A known customer development that never reached finance calls for better communication. An unexpected market shift may require updated assumptions without indicating that either automation or human review failed.
Add human review only where it addresses an identified problem rather than routinely duplicating work that a validated process already performs.
APQC’s 75th percentile benchmark figure of 30% is neither a ceiling nor a prescribed target. Some companies may have substantial room to automate data collection; others may gain more from improving the information behind their assumptions. The benchmark alone cannot determine the appropriate balance, but it does provide an opening to begin forecast-process discussions within the organization.
Faster forecasting should give leaders more time to respond to changing conditions. As automation reduces preparation work, CFOs should direct that capacity toward questions that could change spending, staffing or investment decisions. The payoff comes when finance explains a shift in the outlook early enough for the business to act.
