Many companies building their own RPA programs have already discovered the difficulties of scaling automation. The investment of support required to maintain digital workforces is often under-estimated, a problem which diverts resources that are critical to scaling the program.
A recent webinar hosted by IRPA (Institute of Robotic Procession Automation), ChoiceWORX and PwC, “Scaling in the Hyperautomation Era”, polled the expert panel and attendees to highlight some of the primary reasons that RPA initiatives are stalling. Watch the webinar here.
13% cited ‘cost of maintaining robots’.
Whether a company is new to RPA or not, the financial impact of the “learning curve” as well as expected robot maintenance is often under-estimated or overlooked. The typical lifecycle of a robot is 18 to 24 months. This expected maintenance should account for roughly 15% of the automation project budget. However, what is more difficult to calculate is the impact of the RPA learning curve. Without a solid robot framework from the start, RPA programs are losing momentum, or even plateauing, as engineers continually take robots out of production to be reconfigured. The automation Center of Excellence should determine, from the start, how they will track RPA’s cost to the business, including licensing, infrastructure, support personnel, and time spent on maintenance and troubleshooting. Learn more about calculating RPA’s total cost of ownership.
24% cited ‘difficulty in selecting processes’.
What is automated is equally as important as how it is automated. It’s tempting to start with processes that cause employees the most strain – and as a rule, this is usually a good place to start. But are the tasks data-driven? Do they rely on structured or unstructured data? Are human decisions required? All these factors are important to consider when selecting processes to automate. The best processes for automation are stable, rules-based and repetitive, have a high volume or fluctuate in volume, and are performed by more than one full-time employee. Learn more about selecting processes for automation.
32% cited ‘creating bots is more complicated than anticipated’.
Members of the C-suite may have unrealistic expectations with regards to the difficulty of robot configuration and integration into the workforce. The hype surrounding automation promises simplicity and accessibility for all users, though that is rarely the case when automating the majority of business processes.
As mentioned before, it’s not just what is selected for automation, but how it is automated. Robot frameworks must be built for reusability, recoverability, and scalability.
- Reusability – Designing for reusability requires adequate evaluation of the processes being automated to break them into reusable chunks that can be used across multiple robots. Doing so will drive quality and consistency across your robotic workforce, as well as enable faster implementations. It’s also a lot easier to update one code component used by 15 robots than it is to update 15 robots separately.
- Recoverability – Recoverability is the most sophisticated, and arguably most critical, component in robot configuration. Every robot must be built to account for what it expects to encounter, as well as what it should do when it encounters something it doesn’t expect. It’s a common misconception that robots “just know” or that the RPA software will magically handle it, but recoverability happens at the robot and the software levels to ensure a cascading fallout does not occur.
- Scalability – To achieve scalability, the automation COE has to establish and continually reiterate on the overall structure of the program. Define what success looks like, establish a process for incorporating successes back into the framework, and work to repeat the framework for each process, in every department across the organization.
32% cited ‘Benefits are difficult to prove’.
RPA is famous for time and cost savings as a result of automating high-volume, repetitive processes. But, time and cost ROI is merely the beginning for many processes. ROI is often intangible and can be found immediately or downstream. For example, a health plan could avoid financial penalties for inaccuracy in processing, while downstream also experience an uptick in patient satisfaction for insurance claims being processed faster. Processing Finance and Accounting tasks with total accuracy means the business avoids the cost of rework. or a vendor being billed incorrectly. 24/7 processing of home loan applications means lending decisions are delivered to borrowers more quickly, which increases the competitiveness of the lender. Automation of insurance claims saves insurers millions in handling mistakes and missed subrogation opportunities. Moving critical business processes can keep operations running while your workforce transitions to work-from-home during a pandemic.
It is the responsibility of the COE to work with process owners to establish ROI criteria at the process level. The COE will also set ROI expectations with senior leadership so that the proper metrics are being tracking and reported against for the life of the program.