In a turbulent world where change is the only certainty, the technological advancement that is robotic process automation can carve out some distinct advantages for treasurers. Kevin Grant, Chief International Officer, Hanse Orga Group, explains.
Chief International Officer
We are living in the age of digital transformation. The Internet of Things, mobile, hyper-connectivity, big data, social media, cloud, in-memory computing – all these elements (and more) are equal part commercial threat and opportunity. Which one depends on company outlook.
Either way, they need to be addressed because customer behaviour is changing and established business models are disappearing or morphing into something new in response. There is also the considerable matter of global economic uncertainty bearing down on profitability, to the point where commercial survival is threatened.
For many businesses across the world, digital transformation is, or should be, very high on the agenda. Why? “The greatest danger in times of turbulence is not the turbulence; it is to act with yesterday’s logic,” said US Economist and Philosopher, Peter Drucker. Companies worldwide are challenged by turbulence every day and need suitable tools and robust data that enable them to be competitive and to stay sustainably successful.
Robotic process automation (RPA) has a key role to play.
Rise of the software robot
Forget the images of powerful mechanised arms and futuristic factories; in this context, the robot of RPA is simply defined as the automation of routine processing tasks using rule-based logic. This ‘software’ view of robotics is something many companies are already using, says Kevin Grant, Chief International Officer, Hanse Orga Group.
Indeed, treasurers afforded the luxury of automated compliance with hedge accounting policy and standards will be running forward and reverse exposure analyses on their current hedging portfolios. These ‘simulations’ are entirely within the scope of RPA, says Grant.
Arguably this is a specialist use-case but RPA can encounter a wide set of circumstances and offer an appropriate response. It is being applied not just to the processing of high volumes of data but also being deployed into a broad subset of scenarios and allowing the ‘machine’ to make the decisions, following the logic programmed into it, when it encounters those situations.
This serves to limit the number of exception items that have to be processed by humans. It also serves treasury’s widening scope of interest in corporate financial data and the need for process improvements and efficiencies to deliver better cash management.
Understandably, RPA has become a hot topic in recent times. It is enabling vast improvements in process effectiveness and efficiency, calling upon precisely the kind of forward-thinking logic of which Drucker would approve. The statistics speak for themselves.
According to The Hackett Group’s research, typical finance organisations can cut process costs by 35% by adopting digital technologies, nearly matching the level seen today by world-class organisations.
A library of solutions
The reason RPA is so successful is that if during the process of extracting exceptions for humans to fix, the software tools within RPA are used to fix the problem, machine learning begins to understand both the issue and the fix. It then captures both within the ‘scenario library’ so it knows what to do next time it encounters that situation. Rather than humans continually remedying the same problem, once RPA has learnt and stored the fix, it has removed that exception from the sphere of human intervention.
For RPA to begin the good work there needs to be a level of configuration based on the client’s workflows, processes and controls they wish to put in place. As a system supplier, Hanse Orga Group will create pre-configured systems that are designed to tackle certain tasks. These will have been drawn from its wide industry experience across customer-types and built into the scenario library.
“Customers will already be a step ahead,” says Grant. “They will already have access to a greater capability than they are likely to require to begin with.” When they come across a scenario that is likely to be new to them, the chances are that a subset of scenario rules will already have been created. Through an easy interface, and following training, RPA can then bring that scenario into the client’s ever-expanding active library of fixes. “We call it self-learning but the client is really selecting the fix from the library that suits their needs.”
Building and using robot
Implementation of a user-friendly and pre-configured system, pitched at business people not IT, should be relatively easy. And Hanse Orga Group’s RPA offering is just that, says Grant. Following initial consultation, configuration and training, updating of scenarios thereafter should present few issues.
For SAP users, there is no need to fear tackling Advanced Business Application Programming (ABAP) code to build new scenarios. Hanse Orga Group has already written this into specific functions in a logical way that allows clients to self-serve. The new functions they create, in turn form new scenarios that are logged in the library as fixes. Think in terms of how an Excel cell knows what to do every time it receives new data, says Grant. It has been configured by users based on current requirements and can be used to automatically tackle the same scenario in the future, or be modified as required. It is not difficult!
The current RPA use-case for treasurers relates in particular to cash visibility, notes Grant. This has historically not always been easy for treasurers. Hanse Orga Group is focusing in detail on accounts payable and accounts receivable and allowing the treasurer to implement efficient processes which are automated to a high degree. This, he explains, takes head-count out of these processes, enabling the focus to shift onto higher value-added tasks.
“Our aim is to build robotic solutions with an element of AI across different financial workflows.”
At a more granular level, data can be processed much faster and to a higher level of accuracy. The information being presented to treasury for decision-making is therefore far richer, notes Grant. The explicit understanding of cash positions, for example, can lead to a reduction of external volumes of cash, or, for those long on cash, the confidence to invest more profitably.
“We are doing all the processing at a lower level on DSO, DPO and DIO,” he says. This allows treasurers to have a working capital management solution that takes into account invoice level data, not just data at the reported bank balance level. Taking this with forecasting information, overlaying it with details of any bank credit facility and actual financial transaction data, a far more intuitive and all-encompassing working capital management solution is created.
A core development from this wide-ranging view, and the data that makes it possible, is the move towards analytics. Treasurers can be led to decisions quicker and the availability of key metrics means KPIs can be put in place to drive further process improvements.
If some of this sounds futuristic, the next wave of RPA is upon us already. Automated decision-making based on the results of analytics and artificial intelligence is possible, notes Grant. This is currently practiced in the area of foreign exchange risk management. A system can connect with a trading portal, making requests for an order to be fulfilled and lodging a competitive bid, and then executing that deal before automatically creating trades in the treasury deal portfolio, updating the cash forecasting within SAP and even feeding into the hedge accounting and risk management areas.
Can it go further? It can, says Grant. “But treasury needs to decide what level of AI it is willing to accept.” This amounts to a definition of rules and risk appetite, he notes. A treasurer may be comfortable with the idea of automating processes or even deal execution. But at what point do they want a process to become an exception item with human intervention? And to what extent should intervention be used before the follow-on action is confirmed? This is a matter of choice and confidence in the technology. There is a curious point to make about the latter.
RPA can process far higher volumes of data at far greater speed than any human. Whether matching invoices or making trades, RPA can look at source data and many forms of complementary data. For example, stock price volatility or the volume of credit default swaps being processed against a set of customers can inform a decision to reduce exposure to a sector or individual financial counterparty. Even meteorological data can be analysed to inform trade decisions. “The fact is that RPA gives a far wider vision, building intelligence into decision-making because ultimately, it is about data management.”
The circuit breaker
The threat of automation to some livelihoods is real. Is RPA’s move into analytics-based decision-making the death-knell for treasury? In much the same way that pilotless passenger planes are a long way off, leaving key financial decisions to a ‘robot’ is unlikely, says Grant. It is a matter of confidence and this can be managed.
“Companies may wish to put circuit-breakers into their automated processes,” says Grant. “In making a call on risk tolerance and putting in those circuit-breakers, they will achieve the level of control they require relative to the level of automation to which they aspire.”
Any RPA system must be capable of creating an exception when given data outside a certain tolerance, effectively pulling the plug on itself. In the treasury space, a statement may reveal a customer is going outside of its payment terms. Tolerance to this can be set, triggering automated dunning at that point. But that trigger point is not always clear cut. Dunning may be best served with a harder or softer approach, according to the value or volume of trade, for example.
“Treasury may have disparate systems but what can be created with them is a unified workflow process in a master SAP system.”
Interfacing this process with the sales system, allowing modification of the credit appetite accordingly, would serve to protect relationships. If a customer always pays 30 days late, then the credit risk is minimal. Payment outside of normal behaviour may be a sign of increasing credit risk. This is where AI tools can be deployed to spot patterns over long periods, just as credit card systems track trends in cardholder spend for anti-fraud purposes.
Taking the decision on where to raise the exception and how to handle it will yield different results. It may require human intervention, or automatic reduction of a credit facility for example. The key is to have access to a range of data to inform that decision.
“Our aim is to build robotic solutions with an element of AI across different financial workflows,” explains Grant. To complement Hanse Orga Group’s own systems it has acquired technology in the past year through Soplex, Dolphin, e5 and Tembit. By using standard SAP treasury functionality and integrating with the AP, payments factory, in-house banking and cash management, credit and receivables management technologies from its acquisitions, it is forging links between historically disconnected treasury and finance processes.
When solutions are embedded within the same technology, in this case in SAP, the workflow follows a natural path, even within multiple incidences of the ERP, says Grant. But workflows from outside this environment, from other ERP providers, can be brought on board and integrated using standard connectivity too. “Treasury may have disparate systems but what can be created with them is a unified workflow process in a master SAP system.”
Big data is the key
From localised data capture and processing, treasury gains centralised visibility and control over every aspect of payments. But, as Grant notes, the point of RPA is that with solutions incorporated within an enterprise data processing environment, “there is no reason why treasurers can’t automate on a far wider set of data points”.
This makes the concept of capturing and exploiting big data – including the new notion of the unstructured ‘data lake’ – central to RPA. Using technology embedded in the same environment, such as SAP, provides a single source of truth from that data. This makes for compelling information and intelligence for treasurers. Greater levels of automation are but a step away.
We are indeed living in the age of digital transformation. With RPA’s help, the vast range of data from which treasury is constituted – past, present and future – suddenly seems like an opportunity after all.