According to recent Accenture research, the majority of CFOs are evolving from chief-accountant to strategic enabler and key partner to the CEO.
As part of this, more than 80 percent of CFOs see identifying and targeting areas of new value as one of their main responsibilities. In addition to helping identify new sources of value and risk, AI can help CFOs establish their own roles as strategic enablers.
Here are three ways in which smart CFOs are using AI to predict business risks and optimize business and investment decisions.
1. Enhanced decision-making
CFOs are called upon to advise on complex, high-impact decisions such as divestments, which need to be transparently justified for financial analysts and investors. These decisions are based on numerous metrics that do not always reflect the complex relationships between different performance drivers.
AI cuts through this complexity by creating data networks that surface correlations between all the potential factors influencing performance, enabling CFOs to accurately model the impact of a decision.
For example, we recently worked with a retailer facing a profitability crunch. It knew it needed to close some stores, but it wasn’t clear which would be the best to retain. This decision would traditionally be made based on performance metrics for individual stores combined with general demographics analyses.
However, AI enabled the company to understand how much customer value it would be able to transfer from one store, if closed, to others in its vicinity or to the company’s e-commerce channel.
We were able to map out a network of store-to-store proximity and then develop an algorithm that, for each one, predicted whether customers would be likely to move to one or more stores in the vicinity if their preferred store was closed. The algorithm—for which we were recently awarded a patent—took into account whether the customer was a long-term client buying exclusively from that store or from multiple stores (and therefore motivated to find another one) or a new client who would more likely move to a conveniently located competitor.
The model enabled our client to see whether keeping a store open, despite low profitability, might still make sense in order to prevent customers from going to competitors nearby. This would not have been possible with the traditional method of decision-making.
Not only was the company able to use the algorithm to inform their decision on which stores to close—and to give investors more confidence in their rationale—it can now track data on what customers actually did when the stores closed. This customer retention data can then be used to further optimize the prediction model.
2. Early-warning systems
For companies in highly-cyclical industries, and those influenced by fluctuating commodity prices, accurate forecasting of economic indicators such as GDP, consumer price index or housing starts indicator can be the difference between success and failure. These leading indicators are published only a month, and in some cases a quarter, in advance, giving businesses very little time to react and adjust their strategies.
Therefore, more and more CFOs are looking for ways to get these indicators as early as possible to help anticipate economic environment changes and enable proactive decisions about risk exposure, capital investments, asset sales and activity forecasts. By combining external data with a company’s internal data, AI can generate monthly or even weekly predictions and generate alerts about potential changes in the economy way before those leading indicators are published.
For example, companies can use this approach to build an early warning system. They can learn from the changes that took place in different business cycles and their impact on sales, production and other areas by using internal data from different business units on components such as orders, sales, inventory, production and labor data. They can then use machine learning algorithms to build nowcasting models (which make very short-term predictions) to understand the causal relationship between these internal factors and macroeconomic indicators.
Ultimately, this can help companies identify a set of early warning signals which can be predicted regularly and be made available before the standard indicators. These signals can alert the CFO to potential future threats and opportunities that could affect the business.
3. Tailored forecasting
Organizations can also enhance their investments by using AI to find patterns in the way their unique internal data changes with external trends. These approaches require AI models to be trained on vast amounts of information, which can often uncover unexpected relationships that a CFO would never see without AI.
For instance, if you’re a CFO who wants to improve how the business optimizes up- or down-scaling of investments across a broad range of assets, we can build an algorithm that establishes a correlation between variables, including sales figures, inventory and accounts receivable, and shifts in commodity prices. The algorithm can then discover, for example, that oil price movements correlated strongly with sales of a particular polymer. And when sales of that polymer start to dip, the CFO can know it’s time to shift its investments in oil futures.
AI expands the CFO’s powers
The examples above demonstrate the power of AI to help CFOs make sounder decisions, including on strategically-critical matters. AI is able to make sense of wide-ranging swathes of external and internal data that enable CFOs to make decisions on a broader range of business questions than was previously possible. And, like in the example of using AI to understand customer behavior when the company closed its stores, the insights generated—and the analyses behind them—can be relevant for other business decisions. Such AI-enabled capabilities have the potential not only to enhance CFOs strategic value alongside CEOs but make themselves indispensable advisors to the rest of the business.