Scaling RPA: before automating processes, improve themMost enterprises aren't scaling RPA across their entire organisation, in part, because they don't understand their processes to begin with
When we’re at work going about our day-to-day tasks, it’s not often we stop and think about how our role came into existence, or why we do it the way that we do.
But it’s rather interesting; in fact, the way that we work today is the result of centuries of thinking about how best we should.
From Adam Smith’s “division of labour” to Motorola’s “Six Sigma“, our work processes have been subjected to countless revisions.
New labour laws, economic factors and ever-evolving technologies also mean our processes need to function in a constant state of flux and adapt when necessary.
Robotic process automation (RPA) is one of the latest change agents redefining our work-related processes. While there’s a lot of hype around RPA, its benefits–if applied correctly–are very exciting. Primarily, RPA promises to digitise rudimentary rules-based processes allowing employees to focus on more complex tasks; naturally, it’s become a real boon for back-office operations. As such, interest from enterprises is growing exponentially. In 2018, the global RPA market hit $1.7bn, according to figures from HFS Research.
Failure to launch
However, HFS Research, in conjunction with KPMG, carried out another study to understand how RPA engagements among enterprises are actually shaping up: it’s not all rainbows and unicorns.
According to HFS, despite the billion-dollar-plus vendor valuations, RPA is still a nascent market and adoption strategies are failing to reach fruition, with only 13% of enterprise RPA initiatives achieving scale across the entire organisation.
The issue around scaling RPA is multi-faceted; organisations are struggling with a dearth of experienced talent; they are failing to understand the technologies needed and they are failing to asses processes before they’re automated.
Robotic Process Automation in 2019: the market will come of age says Kofax
The blame game
Perhaps, RPA vendors have been too focused on demand as opposed to supply; also once service providers leave, enterprises are left to fend for themselves, often with little in-house talent.
Obviously, RPA vendors want to advertise the benefits of their solutions, but one can’t help but notice that certain mythologies have arisen. One big piece of hype around RPA right now, according to Elena Christopher–research vice president, industry research and RPA at HFS Research–is time to benefit.
“While, yes you can get benefits from RPA in the short term, it can also become a bit of a catch-22,” explained Christopher. “If all vendors and enterprises are doing is just bringing in the automation and slapping it on manual processes and not really changing the process, sure, you can get that up and running pretty quick but it may not be what you want at all. If, on the other hand, you take a little bit longer to be thoughtful around what you want to achieve you’ll get a better return on investment.”
RPA: the key players, and what’s unique about them
Vendors may also be taking too much of a standardised approach, thinking in the realm of pushing automation on processes they’re usually good at. Instead, vendors should be thinking about problems unique to their clients.
However, according to Christopher, when it comes to picking what processes to automate, enterprises themselves need to be in the driving seat.
“It’s the enterprise, not the vendor, who knows the most about the processes,” she explained. “So they’re really the ones that have to judge whether a process is in good shape or if there is a problem with it.
“A service provider can definitely bring the right tools and impetus to help with the identification but it’s the enterprise that really has to step up to the table and identify what processes are the right candidate for automation.”
According to Christopher, it appears that in the rush to adopt RPA, enterprises may not be taking an integrated approach to automation and are failing to comprehensively examine processes before they automate them.
“I’ve heard lots of enterprises actually admit this,” said Christopher. “Before one of our recent roundtables on intelligent automation, I was going around the room talking to different business leaders and lots of them admitted how when they were starting out they’d look at processes and say ‘let’s just swap in some automation’, then they realise they’re just left with a different form of a worker doing the same job.”
It seems enterprises are leading with a solution before identifying the problem. “Automation should be seen as an opportunity to drive dramatic process improvement,” added Christopher.
This, of course, is no mean feat. Let’s say, for example, you’re a multi-national producer, and you want to improve your order to cash process. From order entry all the way through to the delivery of goods and receipt of payments, it’s a huge project. Before you automate you need to find out a lot; where are the throughput times between delivery and invoicing too high and why? At which steps is manual and time intensive rework required? How often do unnecessary delivery blocks and order cancellations occur?
Speaking with Information Age, Alex Rinke, CEO of Celonis, the process mining technology firm, shared some of his insights; his firm helps enterprises tackle bad processes. In his time, he’s heard of many failed RPA initiatives; and it’s often because enterprises fail to understand their inefficiencies before implementing automation.
Process, technology, and team: the core components of data management
According to Rinke, process mining, a type of data mining that identifies patterns in event log data, can help enterprises take a more strategic approach to automation. “It really helps visualise activities,” he said. “It helps to understand how processes are correlating; it creates a map of what employees are doing and where they are getting stuck.”
By assessing the processes, Rinke argued enterprises are able to balance the pros and cons of automation; and better move to developing bots.
“Our client Vodafone has been able to increase automation in their accounting department successfully,” claimed Rinke.
With 446 million customers, Vodafone is one of the biggest telecommunications companies in the world and, naturally, it deals with hundreds of thousands of transactional data. With process mining, Vodafone has had some big wins, under its ‘touchless invoice automation’ initiative, their invoice automation rate has increased from roughly 13% to approximately 40% (figures from March 2018).
“You can’t just achieve this with RPA alone, you have to really look at what’s going on and then imply the necessary measures in a strategic way,” explained Rinke.
However, while RPA vendors and partners such as Celonis can provide helpful tools and pre-configured apps, to truly understand processes, enterprises are going to have to develop and apply their own expertise too.
Process interpreting and event collection requires technical skill sets, so organisations are really going to need to get their data expertise involved.
Staffing RPA appropriately and getting some commitment behind a data and intelligence-driven approach to process information and to change is also very important. You need to gather them and say ‘this is a part of your job now’ otherwise, you will struggle when it comes to scaling RPA successfully.