With business slowly returning to normal, you might soon find yourself discussing AI and machine learning with your work colleagues at meetings or over coffee again. After all, these innovations are becoming ever more present in our daily lives. A simple Google search, for example, uses multiple machine learning models to define what results we see and in which order, based on data like our age, location, past searches, likes and dislikes, and more.
But how does machine learning and AI fit into the data-focused, analytically-driven world of financial planning? Keep reading for an overview of machine learning, how it applies to FP&A, and the challenges it still faces.
The link between AI and FP&A
CFOs are constantly on the lookout for ways to improve their company’s performance and drive competitive advantage. And with all the hype surrounding AI and machine learning, it makes sense that finance professionals are looking to these new technologies for ways to streamline their processes and get richer insight into their company’s financial and operational performance.
Unfortunately, AI still isn’t built into most financial planning suites. In fact, most AI applications we’re familiar with come from outside the finance world. Some common examples of machine learning and AI we use daily include:
- Virtual personal assistants like Siri, Alexa, and Google Now.
- Face recognition in social media.
- Email spam filtering.
- Search engine result refining.
- Tracking ads.
However, machine learning could become more involved in finance in the future. After all, FP&A and AI share a common thread; both are driven by comprehensive qualitative and quantitative analysis and future planning/predictions.
Some ways CFOs and other finance experts expect AI technologies to transform FP&A us through their ability to run scrupulous analysis of complex data sets while removing human error and automating typically labor-intensive analytical tasks.
Machine learning and AI also promise to help get us out of the trenches of analyzing and reporting on data, instead allowing us to focus on activities that require human judgment, expertise, and decision making to drive real value for our organization.
The gap between FP&A and AI
So, there’s clearly a lot of potential for machine learning in FP&A. Unfortunately, we’re yet to see that potential be exploited.
Right now, most of the applications for machine learning we’re seeing in the finance world are focused on automating financial operations. These include:
- Transaction validation. Support teams around the world spend hours verifying transactions every day. Machine learning algorithms can map the specific rules needed to complete a transaction and thereby automate a big part of this job, letting support teams focus on more valuable ways to assist their customers.
- Trading patterns. In trading, AI can be used to create profiles based on a trader’s behavior, provide recommendations on how to proceed under certain market fluctuations, as well as predict a position’s opening and closing prices, taking into account the specifics of a market as well as that trader’s specific trading patterns.
- Fraud detection. Just like machine learning can help validate transactions, it can also help spot suspicious financial activity. Data science algorithms can analyze transaction data from banks, for example, and automatically create alerts when they encounter suspicious activity. Machine learning systems can also be used to build new detection models based on a customer’s spending patterns.
Possible future AI applications in FP&A
As we saw earlier, AI and machine learning have a lot in common with financial planning and analytics. The fact that these technologies are so heavily based on analysis and prediction means they hold a lot of potential for automating and streamlining FP&A processes.
“I see the most potential for machine learning in risk management,” says Apliqo CTO Scott Wiltshire. “Machine learning could, for example, completely automate Monte Carlo analysis by flexing key driver variables like exchange rate fluctuations and FX exposures, cost-saving forecasts, and market growth rate projections.”
However, in order for machine learning and AI to leave a real mark on FP&A requires fully-integrated systems and well-defined data models. In order to be successful, these systems need to solve real-life problems facing a company. Right now, for example, some of the biggest challenges facing any FP&A software are:
- Providing adequate insight into complex business and economic circumstances. Your team needs to have a clear understanding of how changes to the global economic climate might impact the key drivers of your business. This goes beyond putting together yearly budgets and balance sheets and instead calls for a deep analysis of how changes in global economics will impact every aspect of your company’s operations.
- Dealing with uncertainty. In one of our recent webinars, Jack Alexander confessed that in his 40+ year career in finance, he has never seen a time of greater economic uncertainty than right now. Unfortunately, uncertainty will always be part of financial planning and analysis, and your planning software needs to present you with the right tools and analytical concepts to prepare for this uncertainty as best as possible.
- Cash flow management. Cash is the lifeblood of any business. In these uncertain times, properly managing cash and working capital is becoming increasingly challenging and ever more crucial for all businesses.
Unfortunately, using AI to solve these problems is extremely difficult given the lack of integrated systems and data models. The level of integrated, well-structured data that most FP&A teams are working with simply isn’t deep enough to be applicable for machine learning and probabilistic models.
For more insight into some of the key finance tools to help you deal with these challenges, check out our recent webinar on Key Analytical Tools and Concepts with Jack Alexander. Also, stay tuned for our upcoming article on Scenario Planning, one of the leading techniques for getting insights about the future and dealing with uncertainty.