Dear finance executive,
AI forecasting is the #1 digital use case in Finance, in my view.
Forecasting is a top priority for most CFOs. With ever-increasing uncertainty, looking ahead more frequently has become essential (more detail on this here).
No wonder, it ranks #3 among the most impactful AI/GenAI finance use cases in our recent study with 280+ CFOs.
AI forecasting will significantly up your game in three levels.
Level 1: Better, faster, cheaper
It’s better
Multiple examples prove that algorithms are much better at forecasting than humans, especially in the short term up to 12 months, where there is a lot of data to draw from. Algorithms help you avoid sandbagging, politics, wishful thinking, sunk cost fallacy, or anchoring and confirmation biases.
Unilever improved its demand forecasting accuracy by 20% using algorithms. Daimler FS improved P&L forecasting quality by 30%. ASLM increased forecasting for hedging purposes from 70 to 95%. Further examples abound.
Managers need to accept that algos deliver results that are better - or at least comparable - to those of humans. Don’t believe it? Test it for yourself!It’s faster
Even sophisticated forecasting algorithms run in hours, not days. This allows you to update your forecast almost as often as you want. While a monthly frequency usually works well for a company-level forecast, there might be significant events that would justify an intermediate update. And there are other forecasts - in sales and operations - where daily updates can produce better decisions.It’s cheaper
Running a monthly forecasting process manually is typically very time-consuming. Algorithms automate that effort.
It’s important to note that - despite saving significant effort in finance - it typically does not lead to FTE reductions. Forecasting is a distributed activity taking shares of people’s time, which makes it hard to find full FTE savings. In addition, many CFOs choose to reinvest the freed time of their controllers into more value-adding activities in performance analysis and business partnering.
Level 2: Tangible business impact
Having frequent, more precise forecasts has direct implications for your business financials, for example:
Revenue and margin: Better demand forecasts enable dynamic pricing, promotion planning, and sales allocation. Better cost forecasts allow you to improve project pricing.
Working capital: With better demand forecasts, you can lower excess inventory and reduce out-of-stock issues.
Operational costs: Better demand forecasts improve your capacity utilization, ramping up staff and machinery only when needed and better timing of maintenance during downtimes.
Cash and financing costs: You can reduce your excess cash levels - investing your short-term excess cash to increase interest income, potentially lower your financing needs and costs. Stronger P&L forecasting can also reduce your hedging costs.
The impact of better forecasts on the business can be substantial and goes far beyond the efficiency impact within finance.
Level 3: Upgrading performance discussions
Typical performance discussions center on actual vs. budget comparisons. If you think about it, this actually compares the past (month) with a past plan. No wonder, it creates a backward-centric discussion with accusations and justifications.
Once you have an unbiased, algorithmic forecast, you can go further. You can move to a forecast vs. forecast model:
How has our year-end forecasts changed vs. last month’s forecast?
What do we need to change so next-month’s forecast improves?
This upgrades the performance discussion - it becomes more forward-looking and more action-oriented. The AI forecast acts as an unbiased feedback on past decisions, an impulse to change, and a performance bar that the business needs to drive up.
Practical takeaway: The best time to start with AI forecasting was yesterday, the next best is today.
All the best,
Sebastian