On the road to the intelligent enterprise: AI is opening the way for finance offices to automate a host of procedural tasks – and changing the traditional role of AR accountants in the process. A visit to SAP’s Shared Service Center in Singapore.
Imagine traveling at a snail’s pace and you’ll have a pretty good idea of the speed of progress in Manila’s daily traffic chaos.
When SAP wants to bill a customer in the Philippine capital for the use of its software, it hires a courier to deliver the invoice to the customer by moped on a specific day and at a specific time. Once the customer is ready to settle the bill, SAP hires another courier to collect the customer’s check by moped and return it to SAP.
Hearing Thomas Zipperle, CFO of SAP South East Asia, speak of the everyday challenges he faces, you begin to realize that while the entire world may well be talking about digital transformation, there are still many areas where, for all manner of reasons – from regulatory through cultural to technical – it is still far from a reality. After all, even SAP customers find it hard at times to keep up with the pace of innovation set by their software provider.
Nevertheless, Zipperle is convinced that digitization is an unstoppable force that will “dramatically change the way we work in the future”. Finance, he says, is one of the areas that will be impacted most.
Finance processes in the shared service center
A look behind the scenes of Global Finance Shared Services (GFSS), SAP’s finance shared service center (SSC) organization, proves his point. More than 1,500 employees at five locations across a global network are responsible for handling SAP’s financial processes, including “order to invoice”, “procure to pay”, “cash collections”, “travel and expenses”, and “record to report”.
In 2017, some 2.9 million documents flowed through SAP’s ERP system for the “record-to-report” process alone. These related, among others, to revenue postings, employee compensation, fixed assets accounting, monthly, quarterly and year-end closing, and invoice clearing in accounts receivable.
The SSC in Singapore, which is responsible for the Asia Pacific Japan region, issued around 230,000 invoices to customers and partners last year. In the process, SAP’s AR accountants had to overcome all manner of minor and not-so-minor obstacles.
Let’s look at what happens when the moped courier delivers a check to SAP. It must first be taken to the bank to be cashed. So far, so good. But sometimes, the invoice number is missing from the check and hence from the bank statement. Or discrepancies arise because the invoice currency is different from the payment currency. Or the amount of the payment does not tally with SAP’s invoice because the customer has only made a partial payment or combined multiple invoices in one payment.
All these scenarios mean that the SSC employees spend a great deal of time matching bank statements with invoices. “But that’s all changing now,” says Anja Langhoff, who is head of the GFSS in Singapore.
Pilot customer for SAP Cash Application
She and her colleagues began working with the SAP Cash Application solution in October of last year. Using SAP Leonardo Machine Learning capabilities, the solution suggests which paid amounts might match a specific invoice. “If the customer has paid a single invoice, the hit rate for suggestions is already very high,” explains Langhoff. “It’s not quite so good in the so-called “multi-invoice scenario” – where a customer groups several invoices in one payment – but the system is learning with every passing day,” she adds, “and we are very optimistic that we will soon be able to significantly reduce the amount of manual effort required in this process.”
Employees are already feeling the benefits of the solution. “In a nutshell, it cuts out the tedious, repetitive, and time-consuming tasks and enables us to take direct action based on the proposals,” says Jia An He from the SSC team. “That has a positive effect on motivation, because we gain time for tasks that are strategically important and, by definition, more valuable to the company.”
Which is good news for Anja Langhoff and for Thomas Zipperle, who – like any other CFO – is tasked with helping drive the company’s value and efficiency. “Shared service centers used to focus mainly on transactions and recurring processes,” says Zipperle. “Today, our SSC personnel are more frequently handling tasks like analysis and planning; we have teams that are responsible for controlling governance and for developing guidelines for cost accounting.”
Alongside invoice approval, Zipperle is convinced that artificial intelligence and machine learning will lead to more and more finance processes becoming automated in the future, including expense claim auditing, risk assessment, and approval workflows.
On the road to the intelligent enterprise
For SAP, too, the automation of parts of the record-to-report process is just one of many stages on the journey to the intelligent enterprise, as Sebastian Schroetel explains. He is in charge of development for machine learning solutions around SAP’s digital core and has a clear vision of where corporate finance is headed. “Today, it’s users who manage a business process. They open a document, make a decision, close the document, and forward it to the next user. In the future, a process like this will no longer be controlled by a person, but by the system and by artificial intelligence. At SAP, we refer to this as “intelligent process automation”.
Take the example of month-end or year-end closing: In an intelligent enterprise, says Schroetel, it is the software that will initiate and manage these processes. If the system encounters something it can’t decide about itself – such as which depreciation rule to apply – it will involve the employee it considers most competent to make that decision. The answer supplied by that employee will help the software learn and improve its capabilities. “Business users will gain a new identity and self-image,” says Schroetel. “They will no longer be executing the same processes over and over; instead, their skills will be required for strategic decision-making and problem-solving.”
Intelligent process automation
Back in 2016, consulting firm McKinsey estimated that around 70% of all financial tasks could be automated. According to Schroetel, financial operations and period-end closing are particularly well suited to digital automation and the use of artificial intelligence. “Finance is number-driven; it generates a wealth of data; and it involves a large number of standardized processes.” Processes like those running in SAP’s shared service centers.
In terms of the technological framework, SAP is relying on SAP Cloud Platform and the continuing augmentation of the SAP Leonardo Machine Learning Foundation. This offering provides the technical services to implement machine learning on SAP Cloud Platform. For example, the ML Foundation already delivers the components required to match invoices with payments in the SAP Cash Application solution.
SAP will deliver new functions every quarter and make it possible to integrate additional information into the SAP Cash Application and other machine learning solutions. For instance, SAP will soon also offer a business service for accounts payable on the ML Foundation. And a service for managing lockbox data – a prominent feature of payment transactions in the United States – is slated to be available by SAPPHIRE NOW in June. In addition, SAP is planning to offer intelligent process automation across the boundaries of the ERP system by processing financial operations via autonomous bots.
All of these short and medium-term steps will change the world of finance, automate processes, and bring SAP and its customers closer to the vision of the intelligent enterprise. And, inevitably, they will change the task profiles of the people who work in those enterprises – whether they operate in a traditional accounts department or in a shared service center. But, Schroetel is convinced, “Their contribution will become more valuable”.