Robotic process automation (RPA) promises to replace expensive, error-prone humans performing standardized, repetitive work with software “robots” that work faster, at lower cost and without error, and deliver an impressive ROI. Better yet, the robots are perfectly content to perform tasks that their human counterparts find mundane.
Many companies, however, are having issues with the “ROI,” “standardized,” and “repetitive” parts of that definition. System integrators and platform vendors tell us that 50% to 70% of implementations fail to advance past the pilot phase. Why? In part because companies fail to understand the implications of those three terms as they relate to RPA.
What are those implications, and what should be in place in order to replace human labor with software robots? Just how standardized does a task have to be before it can be automated? At what point is a task repetitive enough that the cost/benefit calculation favors automation?
Those are the key questions you need to consider when planning your RPA journey. The pointers below can help keep you from straying off-course.
RPA has many different business justifications — improved customer service, reduced error rates, faster cycle times — but if you seek a hard-dollar ROI (cost reduction, not cost avoidance), you will likely have to reduce headcount. That might be difficult. Consider a company that wants to implement RPA via its IT department. IT has already succeeded in automating some processes, tasks, and activities. But the headcount hasn’t budged, because the people whose tasks have been assumed by robots perform other tasks as well. We can express this as a simple arithmetic exercise: A 100-person department that works 2,000 hours a year represents 200,000 hours of work time per year. Automating 30% of that work (60,000 hours) is not the same as automating 30% of the people.
The moral of the story is that automation alone is not sufficient to realize headcount reductions. Re-organization, active change management and process redesign must go hand-in-hand with automation if hard-dollar ROI is to be realized in this case.
If, on the other hand, cost reduction is not your key goal for RPA, your CFO will still thank you if you can draw a straight line from your RPA investment to desired financial outcomes.
Consider another company that is embarking upon its own RPA journey with a project to automate an already-standardized, if relatively trivial, task (one that the humans are delighted all the same not to have to perform). Problem is, the task isn’t as standardized as it first appeared. It involves a single (but complex) spreadsheet form, email and an accounting application. The application, however, actually requires three separate forms to cover special cases. Those “special cases” vastly complicate the work of automation and can stall RPA implementations that at first appear relatively simple and straightforward.
RPA implementation relies on mapping processes, tasks, and activities in exhaustive detail, all the way down to the keystroke level. For example, a task involving 15 separate steps on a flow chart may have many times more elements to it when broken down to the detailed keystroke level. Most organizations are not prepared to plot that breakdown themselves and will need to engage an implementation partner with the skills and patience to map tasks at that level of detail.
A related issue is that perennial bugbear of IT departments worldwide: data standardization and cleanliness. Proper data hygiene is still a stretch for many companies, but it’s vital. Messy data will multiply the number of exceptions to the RPA process that require human intervention.
The more repetitive the process, task or activity, the better candidate for RPA it is and the higher the potential ROI. But just how repetitive does a task need to be?
Some industries, such as financial services, and functions, such as call centers and shared services centers, are obvious candidates for RPA. But less obvious applications will likely require some time to analyze and map. The good news is that as your RPA implementation ramps up and gains momentum, successive automations will happen more quickly, and it will be easier to justify the investment to automate processes, tasks or activities that entail lower levels of repetitiveness. Keep the following questions in mind when deliberating where your RPA journey should take you next:
- What have you already outsourced to service providers? Which activities are ready to realize the next level of benefits from automation? If those activities are highly repetitive, they can support a business case for bringing them back in-house.
- For which processes, tasks or activities do you maintain metrics? These are also often the most repetitive ones.
- Simply asking your staff, “What are your most repetitive tasks?” can open up lines of inquiry that end with the identification of processes, tasks and activities that are good candidates for automation.
Properly planned and executed, RPA works. Successful programs usually have a payback in less than a year and generate a strong hard-dollar ROI. So get started somewhere with RPA. A willingness to proactively address efficiency issues, in parallel with automation initiatives, can start you on the path to automation—if your senior leadership is behind you. Don’t let paralysis stop you from helping your company and workforce through RPA.
Your RPA checklist:
- Embark now on process standardization. You should be doing this anyway, and you may find after doing so that you capture a lot of the benefit that would otherwise have come from automating a less efficient process.
- Commit to a hard-dollar ROI. If you are not committed to reducing headcount, make sure you can draw a line from your investment directly to the ROI opportunity. Think cost reduction, not cost avoidance, and be sure you can clearly attribute a revenue increase to RPA introduction. You need to do this to persuade your company’s leadership to fully embrace RPA.
- Be prepared to reorganize your work to capture the benefits of an RPA program.
- Maintain an active governance structure and pipeline of automation opportunities ranked by their ROI, which should be related to the degree of repetitiveness and standardization.
- Clean up your data, at least make it cleaner than it is today.
- Remember, RPA should not be the first answer to every automation opportunity in your company. It should be last thing you choose once you have exhausted the options of eliminating work, reorganizing work, or finding a better, cheaper, quicker way to automate.
Jim Shand is Director in AlixPartners’ Digital Practice.