Building The Next-Generation Treasury

From AI to robotics to virtual desktop assistants, a host of new technologies promise to make the treasurer’s life easier.

Has there ever been a better time to be a treasurer? Most multinational companies are sitting on pots of liquidity, which could be used for M&A or capital expenditure. And since the 2008 global financial crisis, the treasury and finance function is more strategic than ever to the overall business.

The UK Association of Corporate Treasurers (ACT) Business of Treasury 2018 report finds that almost 90% of treasurers surveyed “believe that treasury now enjoys a strategic position in their organizations.”

The makeup of treasury operations is also changing, according to the ACT survey. “Treasurers don’t look like they used to,” it notes, thanks to generational changes within the profession and new skills that treasurers will need to acquire as some functions become increasingly automated.

While most treasury operations are already heavily invested in enterprise resource planning (ERP) software, which integrates a number of operational functions, these systems have not always delivered the information or insights that treasurers need to better manage their cash and liquidity. New technologies promise to make life a lot easier for treasurers, enabling them to more accurately predict long-term cash and liquidity needs and better manage unforeseen risks. A recent report from Capgemini’s Digital Transformation Institute concludes that “through ‘intelligent automation,’ the right combination of robotics process automation (RPA), artificial intelligence (AI) and business-process optimization…the financial-services industry could expect to add up to $512 billion to global revenues by 2020.”

Moroney, ActiveOps: The majority of companies are doing [robotics] cautiously; there is a significant change-management challenge.

Introducing RPA

Capgemini’s findings pertain mostly to large banks and insurers that are already experimenting with these technologies. However, highly manual and repetitive tasks within the treasury, accounting and finance functions—such as payables, collections and reconciliations—could also potentially be transformed. “A lot of reconciliation activities are being automated right now,” says Sankar Krishnan, executive vice president for banking and capital markets at Capgemini. “Robots can read and write copious amounts of data and cross-check data against systems, which reduces not only costs, but also errors that ultimately cost businesses money.”

Paul Moroney, chief product officer at ActiveOps, which provides back-office workforce-optimization tools, says RPA is best suited to repetitive tasks of short duration. “That’s why we see it in locations like shared service centers [SSCs],” he says, where it can help lower costs and increase reliability. In a report on the future of finance functions, Ernst & Young says that RPA challenges the economics of shared-services offshoring with its “relatively short implementation timelines and low maintenance costs.”

Robots don’t make mistakes, notes Oded Karev, head of advanced process automation at enterprise-software provider NICE. “They are dependable and can help companies meet service-level agreements and enforce compliance.”

Finance functions such as billing and collections, accounts receivable, reconciliation, financial reporting, budgeting and forecasting readily lend themselves to RPA, EY argues. But some of the more specialized and complex functions within treasury, such as liquidity management and managing outlying risks, are less amenable to being performed by a robot, Moroney counters. “The majority of companies are doing [robotics] cautiously,” he notes. “There is a significant change-management challenge for organizations.”

Capgemini’s study finds that “only 10% of companies have implemented the technology to scale,” as the majority of firms still struggle with the “business, technology and people challenges” of doing so.

Treasurers don’t seem daunted by the prospect of machines in the future taking over a lot of the tasks for which they are responsible. “There may be fewer of us in future; but when things go wrong, and a company can’t source liquidity, you can’t hand that over to a machine,” says Stephen Baseby, associate policy and technical director at ACT.

Baseby argues that companies adopt newer technologies the moment their current systems prove inadequate, often because of fragmented IT. “That may well present an opportunity for businesses to make that IT shift,” he says. “But it’s not like smartphones, where everybody is going to rush out and buy the latest one.”

Baseby says that a lot of the newer, more intelligent technologies currently being talked about are more likely to appeal to accounts payable and accounts receivable, which are focused on resolving payment problems. “Treasurers sit on top of that money food chain,” he says. “We can use this information to start refining our own cash-flow forecasting processes.”

Robots Reach the Desktop

Treasury typically runs lean. RPA today is being deployed mostly in departments with a large head count and for less-complex manual processes that lend themselves to automation, says Karev. The next phase of technological innovation, he predicts, will see robots move onto the desktop—and this may have a broader impact on treasury.

In May, NICE rolled out a new virtual desktop assistant, NEVA (NICE Employee Virtual Attendant), which combines robotics with AI. Karev says that with NEVA a normally 14-minute process like fraud analysis can be reduced to a two-minute conversation, which improves service-level agreements and the handling of fraud alerts. The market for such products—intelligent virtual assistants—will top $11 billion by 2024, according to Global Market Insights, a market-research and management consultancy.

Banks are also getting into the act. For its consumer and small-business mobile-banking customers, Bank of America Merrill Lynch developed an AI-driven virtual assistant called Erica, which began its rollout in March. “Erica combines the latest technology in artificial intelligence, predictive analytics and natural language to be a virtual financial assistant to clients,” says the bank. It can search for transactions, view balance information and bills and obtain credit scores. The bank’s Intelligent Receivables solution uses AI and optical character recognition to help companies reconcile incoming payments more quickly.

BofA Merrill says AI can automate repairs for poorly formatted payment instructions and increase straight-through processing, providing greater certainty and quicker finality of payments, which is important for treasurers. The bank anticipates new services being developed using AI in areas such as cash-flow forecasting, more effective use of surplus cash, and just-in-time delivery of working capital.

Karev, NICE: Robots are dependable and can help companies meet service-level agreements.

ERP software also is being more tightly integrated with AI and machine learning, which could help treasurers better predict interest rate fluctuations and liquidity needs, says Krishnan. Vinod Bhaskaran, global solutions lead for trading, risk and regulation at software provider Finastra, notes that an virtual assistant can help treasurers sift through reams of data to more easily and quickly identify where they are likely to experience liquidity shortfalls.

The Tipping Point

Experts are divided on how soon such tools will become commonplace. “All it takes is a few banks and treasurers to start using these technologies, and it will reach a tipping point,” says Bhaskaran. Baseby, however, believes adoption by treasury departments will be more measured. “A virtual desktop assistant for treasurers may be possible in future,” he says. “But what if something goes wrong? As treasurers, we tend to plan on the assumption that things will go wrong somewhere—and they do. Providers will have to overcome that barrier, as treasurers need to have confidence in the software.”

Intelligent automation technologies are more easily adopted by newer businesses—the Ubers and Spotifys of the world—rather than FTSE 100 companies. “You have these established processes out there that you can’t just get rid of,” Baseby says.

Treasury operations at multinational companies are unlikely to be transformed overnight by RPA and AI. The best approach may be for companies to first look at which technologies would be most immediately useful (with least disruption), and make any further transition incrementally. Adoption is most likely to come from intelligent-automation tools embedded within software that treasurers already use, which don’t require significant buy-in or changes by companies’ IT departments. As treasury assumes a more strategic role, the rest will surely follow.