Traditional banks have begun offering businesses of all sizes a tech tool that fintechs had developed for small firms: cash-flow forecasting, or anticipation of how money will flow into and out of a company’s coffers before it’s caught short.
PNC Financial Services Group and Bank of America each rolled out a product last month. JPMorgan and Wells Fargo unveiled their offerings in October. Citi introduced its version early in the year.
The challenges of cash-flow forecasting plague businesses of every size. Many, even large companies, have long turned to spreadsheets to predict their ebbs and flows.
“Your cash position will hit you whether you forecast or not,” said Karl Augenstein, an executive director at the technology and management consultancy Capco. “Knowing whether you will have enough funds for your next project will help you plan whether you need to find a loan.”
But banks have only recently started exploring how they can automate this task, as part of a broader push to use data and insights to engage their customers. They’re incorporating such features into their treasury management systems, using artificial intelligence and machine learning to automate forecasting, or are partnering with fintechs that do the same.
With a few exceptions, “most of the top banks are still talking about just treasury management, liquidity, the historical stuff,” Augenstein said. “‘We’ll help you aggregate funds, we’ll tell you about your positions in foreign currencies around the world.’ But it’s not really forecasting — it’s how much money you have and where it is.”
PNC recently introduced its tool after beginning to mull the idea last year.
In November, PNC’s treasury management division announced the addition of Cash Forecasting to the Pinacle corporate banking platform it’s offered for more than a decade. Clients can upload the data of their choosing, including from enterprise resource-planning systems and other bank accounts. The tool uses artificial intelligence and machine learning to apply one of five models and produce 31-, 60- or 90-day rolling forecasts to predict cash flow, plan for gaps or surpluses and explore current and future cash positions for different scenarios. This product was mostly built internally.
“If you can do cash forecasts more frequently and in a more integrated way, you’re going to be more on top of your business and spend less time keying data into Excel,” said Chris Ward, head of data, digital and innovation for PNC Treasury Management.
The response so far has echoed that goal.
“We’ve gotten feedback from customers about how this has improved their ability to be more decisive about their businesses,” Ward said. “They spend a lot less time on manual tasks and a lot more time on what to do about the forecasts.”
Bank of America launched cash forecasting to its broad customer base last month as well. But the road to incorporating this capability into CashPro, its online banking platform used by business clients, stretches back four years.
“The first thing a treasury client will do in the morning is set their forecast. Before they do their outgoing payables, before they make their decisions on investments for liquidity, they’ll want to know their cash forecast,” said Tom Durkin, global head of CashPro in global transaction services at Bank of America.
Paul Smithwood, product manager for data and artificial intelligence in global transaction services at Bank of America, said the CashPro team repeatedly heard the same concerns from its clients.
“Forecasting is typically an extremely manual task,” he said. “Businesses can spend substantial amounts of time exporting data from bank portals and internal systems, integrating them with whatever tools they have available, which typically are spreadsheets, then trying to do some ad hoc analysis and making liquidity decisions as a result.”
At the same time, they worry about accuracy and having the resources and technical know-how to invest in the task.
After scoping out potential fintech partners and piloting products with a few clients, Bank of America narrowed its choices down to one vendor that specialized in cash forecasting and machine learning and conducted a beta test with 75 clients over the past year. The bank declined to name its vendor.
After customizing the product and gathering feedback, it launched CashPro Forecasting to all clients last month. So far, hundreds have enrolled.
After customers sign up, machine-learning algorithms generate forecasts using two years of the client’s historical data, including balances and transactions within Bank of America, data from other bank accounts and anything else they would like to enter manually. Businesses can create custom categories and assign rules to automatically classify transactions into categories they create, such as payroll or collections. Graphs and grids will illustrate the results so they can quickly interpret the forecast.
Some other major banks have made similar moves.
After a pilot, J.P. Morgan Payments introduced a forecasting tool in October as part of J.P. Morgan Access, its cash management platform. J.P. Morgan clients can see up to three years of historical flows, compare different periods of time and parse through transactions. The feature uses artificial intelligence and machine learning capabilities developed in-house.
In October, Wells Fargo
Citi’s Treasury and Trade solutions incorporated forecasting tools from Cashforce, a forecasting and analytics platform, in the first quarter of 2021. Citi Ventures took an equity stake in Cashforce in 2020.
Fintechs, such as
Bottomline Technologies, a business payments company, offers its own cash-forecasting tools and recently partnered with Centime, a software company that helps small businesses manage and forecast cash flow.
“Cash-flow forecasting is a need that’s been around for a long time, that’s partly been filled here and there, but banks are now taking it more seriously,” said Norm DeLuca, director of banking solutions at Bottomline Technologies.
Augenstein predicts that forecasting is one more way to lock in customer loyalty.
“Once businesses get a taste of being able to do this they will fall in love,” he said. “It’s an incredibly sticky product.”
