25+ successful RPA cases @ Umicore: what’s the secret formula?

In order to thoroughly explore RPA technology and its opportunities, Umicore launched an RPA center of expertise as its first step toward robotics, choosing UiPath as the preferred RPA platform. “We were confident that RPA would quickly deliver business benefits,” Joris Peeters, Director Business Applications IT, begins. “The best way to achieve those benefits was by setting up pilot projects in domains where we expect quick ROI, like treasury, audit, tax and accounting, in addition to purchasing and IT administration.”  

Payback time: 3 to 9 months

“The challenge with projects for finance is that, for us finance teams, the ROI is always top of mind,” admits Koen Van Impe, Global Process Owner Finance. “RPA is a perfect fit, as you can begin with a minimal start-up investment and constantly scale up with new products and robots. Typically, the payback time for an RPA project is around three to nine months. Next to ROI though, data quality has also been a major driver to prioritize cases.”

“However, that’s only true if your processes match with RPA,” Joris adds. “Launching an RPA project just because you want to try RPA is not the right way. Start with the business case: what are the challenges the finance team is facing, what are the processes that need improvement and what processes and workflows are good candidates for RPA. When a process needs reengineering or improving, RPA is often only one of – or part of – the potential solutions.”

Ensuring user buy-in

From experience, Umicore knows that employee buy-in is crucial for the success of any digitalization project. That’s why they closely involved business users in the RPA pilot cases. Koen: “We organized a workshop explaining the basics of robotic process automation and the conditions for a good RPA use case. Together with people from our finance teams worldwide, we then built a pipeline of possible business cases.”

4 conditions for a good RPA use case, according to Umicore:

• The input is digital (or can be digitized) i.e. the data and files are stored on a computer system;
• The input is structured, such as Excel or database files, rather than unstructured (e-mails, input sensitive to interpretation, ...);
• The process is rule based, i.e. it can be visualized by decision trees without too many exceptions;
• The process is stable: it has been executed consistently over a longer period of time (minimum of 6 months).

What processes are ideal for RPA?

Today, Umicore successfully completed over 25 RPA business cases in finance, which can be categorized into three major project types: 

• Replacing repetitive, administrative, rule-based work previously done by humans (e.g the creation of SAP users). 
• Validating the work done by humans against pre-defined rules & policies: tasks like cross-checking documents that involve large data volumes and are therefore tedious and time-consuming to process (e.g. cross-checking all vendor items to flag suspected duplicate pre-payments).  
• Performing high-volume, rule-based tasks and controls that were not or only partially done before because they were too manual or labor intensive: cases mainly linked to data quality and audit processes, like verifying all the coordinates and information of new Umicore suppliers (70 to 80 criteria for 10,000+ suppliers/year).

Human and machine working together

“Robotic systems are ideal stand-ins for people in these cases,” continues Joris. “They are available 24/7 and don’t lose focus, no matter how long they’ve been working.” Koen: “Of course, that sounds quite harsh for people. Some employees were afraid that RPA would take away jobs rather than reducing work. But it is absolutely not our aim to replace people. By eliminating repetitive, error-prone activities, RPA frees up the time of our finance teams to do more interesting, value-added work.” 

“Cross-checking vendor invoices to detect double invoices, for example, is to a large extent, covered by ERP logic, but some cases slip through the embedded controls. By combining RPA with fuzzy logic algorithms, we are able to identify and report suspected duplicate invoices to the A/P accountants. Even if the cases discovered remain exceptions, it’s still a lot of money and it directly impacts the cash flow and P&L. The ROI of this use case is significant, as the solution can also be applied to historical data.” 

Don’t forget the ‘people magical triangle’

While the ROI of RPA is quick and high, RPA tools require quite a bit of expertise to build, integrate and keep running smoothly. “Hence our RPA center of expertise,” says Joris. “More than that, we need three different roles to support an RPA project: technical people with  analytical minds to develop the bots, business people who know the workflows and processes to bring on the ideas and – in defined cases – an IT expert (e.g. an SAP expert) who knows the applications in detail. They needn’t be different persons, but they have to be able to wear different hats.” 

delaware as an expert partner

delaware has been the implementation partner since the early days of RPA at Umicore. Two FTEs are helping the in-house team build and run the RPA technology. “Every use case is a co-development activity,” says Koen. “In addition, we get together every two weeks to discuss the pipeline and progress of RPA cases and discuss possibilities in the future.”

Looking back on 25 successful use cases and with many more in the pipe, Umicore believes in the benefits of RPA. “The power of RPA has been demonstrated. We already started capitalizing on our RPA investments. From now on, every new project helps us further raise (cost) efficiencies and boost our bottom line,” Joris concludes. “By increasingly combining RPA technology with other digital technologies like Webscraping, OCR, artificial intelligence, fuzzy logic, etc. we’ll be able to add even more value in the future.” 

Inspired? Kick off your company’s RPA journey by getting in touch with one of our experts.