Dawn Of The Super Treasurer

Artificial intelligence is moving from the consumer to the business sector, turning the treasury function into a sophisticated analytic center.

Artificial intelligence is creeping into almost all aspects of our lives. Everything from mobile banking apps and chatbots to voice-activated home assistants and self-service checkouts at supermarkets now have some form of AI embedded in them. Given the technology’s pervasiveness in the consumer space, it was only a matter of time before it started to permeate our business lives as well.

At last year’s EuroFinance conference, Spotify shared how it uses different forms of AI, including robotic process automation (RPA) and machine learning, to reduce costs and human error in its back-office treasury functions. Treasury management software vendors like ION see AI and machine learning transforming the humble treasurer into a “super treasurer,” armed with real-time information and business insights that can help them better understand how well their cash, risk and liquidity management strategies are working.

In a 2019 report, Journeys To Treasury, BNP Paribas, SAP, PricewaterhouseCoopers and the European Associations of Corporate Treasurers (EACT) found that better analysis and insights for treasurers often go hand-in-hand with technologies like RPA, AI and application programming interfaces (APIs). The report cites cutting-edge technology companies like Honeywell, which already uses AI and RPA to collate data to better understand their foreign exchange exposures and automate hedging activities.

“This was previously done in a shared service center with transactions undertaken manually, so the project represents a major saving and a streamlined FX risk management approach,” says Rónán Clifford, Treasury Manager at Honeywell.

AI-driven virtual assistants, already standard fare in the consumer banking world, are now populating the staid realms of corporate banking. Following the success of its AI-driven virtual assistant, Erica, which uses predictive analytics and natural language processing to enable mobile-banking customers to more easily search for transactions, view balance information and bills and obtain credit scores, Bank of America Merrill Lynch (BofAML) plans to roll out a similar assistant for users of its CashPro corporate banking portal next year.

Tom Durkin, head of global digital channels at BofAML, says corporate banking customers will be able to use the CashPro Assistant to more quickly and easily find answers to their account queries and questions online. “The assistant will be able to answer questions like how to make a payment from the UK to Singapore,” says Durkin. “Accessing information is top of mind for most treasurers and by using AI, they can be more proactive in terms of how they do this.”

Cash forecasting is another area, says Durkin, where AI could really help treasurers, by harnessing information from all of a company’s banking partners and processing larger quantities of historical data to better predict future cash flows. Software vendors in treasury and cash management are also looking at how the technology could be used to model and test liquidity management scenarios or predict how certain stress events may impact liquidity availability.

Treasurers’ cash forecasting and liquidity management decisions have long been based on intuition and incomplete or inaccurate information. AI provides them with the tools for developing more predictive models that promise a better understanding of the outcomes of their decisions.

“We’re getting to a place where we can now offer tools that enable the treasurer to act faster with increased agility and feel more comfortable with the decisions they are making,” says Gaurav Khanna, head of mobile strategy and new products for global banking at BofAML. Khanna anticipates that predictive analytics powered by AI will find uses in areas such as cash management, forecasting, fraud detection, risk and real-time payments.

Machines may end up doing more of the monotonous number crunching, but treasurers and CFOs will still play a vital role in interpreting data, managing critical exceptions and using information to make better-informed liquidity management decisions. All of these functions will reinforce the more strategic role the treasurer is coming to play within many organizations.

“We’re not going to replace the human touch,” says Khanna. “AI is a complementary function that empowers treasurers to leverage technology a lot more.”

AI Reality Check

Most companies plan to invest in AI in the coming months, according to research by Accenture. However, more than 50% are still in pilot mode or the early phases of adoption while some have not even reached the starting blocks—an indication of the implementation challenges.

“AI can be a bit scary,” says Caroline Hermon, head of AI at SAS UK & Ireland. “But it doesn’t need to be. It’s about identifying the bottlenecks you’re trying to address and then deciding whether AI is the right way to address them.”

Treasurers and CFOs will need to have at least a basic understanding of how the underlying algorithms work; “computer says yes, or no” won’t be enough to justify actions they may have taken off the back of recommendations made by machines. Certainly, regulators will want to see what’s inside the box.

“Traditionally, AI and machine learning have been a black box, which is great at predicting outcomes but limited in providing clarification and reasons,” says Khanna. “But companies need to be able to explain their decisions and ensure that any predictive algorithms they use are unbiased and transparent. If you don’t have explainability in your model, you don’t know what you’re dealing with.”

Scott Edington, CEO of Deep Labs, an AI start-up, says AI has a journey ahead. “You don’t just go from zero to 100 overnight,” he says. “You start introducing some of these technologies in a supervised mode.”

Even for companies that are wedded to legacy technologies like ERP and treasury management systems, Edington says it’s not a case of having to “rip and replace” these systems; AI plugs into existing technologies that companies already use.

The return on investment from AI can thus be a lot quicker than for other technologies, says Hermon, but companies need to prove it in one area of the business before rolling it out to other parts of the organization. “AI is a powerful tool,” she says, “but it comes with some caveats.” Data is critical and needs to be fit for purpose, and although AI may free up a CFO or treasurer to focus on less manually intensive and more value-added tasks, Hermon says they need to think about governance and transparency and put controls in place to measure the value that AI delivers.

“AI is definitely here to stay,” says Hermon. “It’s an efficient way of automating manually intensive tasks and manipulating data, which can be done in a more automated and efficient way.” Referencing Kodak, which failed to predict the impact that digital photography would have on its business, Hermon argues it’s never too early to have a conversation about AI, wherever you’re sitting within an organization. “If you’re not beginning to adopt it, you’ll be left behind,” she says flatly.