CFOs in recent decades have faced several actual or possible nightmare scenarios — the collapse of mortgage finance, a once-in-a-century global pandemic, a sudden worldwide shortage of life-sustaining grains and the risk of the first nuclear weapons attack since 1945.
Most CFOs did not, or would not, predict such crises and the harm to their companies’ revenue and operations, according to Gartner. But by using artificial intelligence (AI) in scenario planning soon after a crisis erupts, a CFO can quickly estimate the business impact and act to limit losses.
“We’re not making a magical prediction about dark horizons,” Mark McDonald, a Gartner senior director analyst, said Monday in an interview. Instead, a CFO using AI can “navigate that horizon.”
By using AI in scenario planning, CFOs can evaluate a much broader range of possible risks and business outcomes and gain greater insights on the probability of occurrence and when and how to respond, McDonald said at the Gartner CFO and Finance Executive Conference.
“AI introduces a very compelling alternative to our traditional scenario planning process,” McDonald said during a presentation at the conference at National Harbor, Maryland. “It’s not going to eliminate your worst case scenario, but it’s maybe going to mitigate a worst-case scenario.”
CFOs tend to shy from gathering a broad range of data when attempting to limit risk through traditional scenario planning, McDonald said. “We’re averse to it in finance because it drives up the complexity and we’ve got to analyze all this new stuff.”
“There’s a need for finance to stop being data-averse and become active about looking for data, not just tolerating it,” he said.
AI-driven scenario planning starts with distinguishing between external influences outside a CFO’s control — such as inflation, unemployment, economic growth and market size — and controllable “internal drivers,” such as staff and inventory levels and spending on advertising, technology and marketing.
A CFO uses AI to analyze data on the external and internal factors that shape an outcome essential to business strategy, such as the future size of a market for a product or service, McDonald said.
The software engages in random sampling, combining the many factors that determine market size into thousands of possible combinations, he said. Within a few hours, AI provides a statistical distribution showing the full range and probability of various outcomes for market size. It does not make a single prediction.
“This computer works nights and weekends, doesn’t take vacation, doesn’t catch COVID,” McDonald said. “We’re talking about seconds, minutes, maybe hours — not days and weeks” as with traditional scenario planning.
CFOs who know the range of highest probability outcomes have essential insights for valuable forecasting and decision-making, he said.
CFOs can use AI “to tell us which decisions to make to pick the best values for those internal drivers that maximize” revenue, he said. “Let the computer tell you exactly which values in which decisions to make, and then continuously monitor your results so that you know exactly when to respond.”