Robots don’t take coffee breaks, and that’s certainly part of their appeal. But they’re often not as busy as they could be, and this means that companies often don’t see the best return on their investments in robotic process automation (RPA). As financial institutions focus on labor utilization rates, a new technique is emerging: keeping digital labor busier by automating more and more smaller tasks.
In the financial services industry, certain kinds of work remained stubbornly manual for years. Now that RPA is able to handle a surprising number of routine activities, perceptions on benefits may not have kept up. We’ve found that some managers still think digital labor won’t apply to their processes because they don’t have big blocks of repetitive work.
Times have changed. RPA can now be used in applications that were impractical, or impossible, just a few years ago. Meet micro-RPA opportunities: defined, repetitive, brief tasks and projects that are now done by regular employees. That report that takes 15 minutes a day to produce and send via email? The five minutes it takes to send reminder escalations when reconciliation breaks aren’t being reviewed promptly? These may seem trivial to automate, but minutes add up.
Such small projects can often be easily configured with today’s digital labor tools. By automating them, firms are able to remove functions from job descriptions where humans don’t add meaningful value. The business case around identifying, automating, and managing these digital processes is compelling.
When we work with our clients to rethink business processes, we typically focus on activity with a lot of associated headcount. We drill into a few high priority opportunities and consider the benefits of redesign. Even with RPA, we find that the best return on investment can come from streamlining first and automating what’s left.
With micro-RPA, however, the process is a little different. We identify a slew of very simple tasks that a given team executes over a period of time. We build and schedule these tasks to run on a bot, and then we re-evaluate the team’s work. Consider a team of 10 people. If a bot can automate around 50 minutes per day of tasks for each person in the group, that’s roughly equal to one staff member. To save those 50 minutes, the bot might perform 20 different activities; “micro-RPA” really means micro. Automating one staff member’s time might not seem like much. But when you consider that the fully loaded cost of an onshore person could be five to ten times that of someone offshore, and that robots can be cheaper still, you’ll see why financial institutions are paying attention. We typically find plenty of small activities that aren’t cost-effective to send offshore but that can be easily assigned to a bot and managed onshore.
How do you find these small opportunities? Here are some examples of staff performing a lot of well-defined, relatively brief tasks:
Micro-RPA projects are different from regular digital labor opportunities. To get the most out of your initiative, here are some things to keep in mind:
You’ll need a different deployment model. Many firms lean too heavily on IT to deploy their initial RPA efforts. This works for large and complex automation processes, but it can be overly complicated and expensive for micro-RPA. We suggest training business users themselves, right from the start, to develop, test, deploy, and monitor the robots. In our work, we’ve seen success by sitting alongside business users and showing them how to program the robots. Once they’re comfortable, they can take on more of the process themselves. This approach also helps you embed the automation culture across the organization. Teams are more likely to embrace technology tools when they’re part of the productivity toolkit rather than having the technology imposed from outside.
You’ll need to measure what you’ve done. When you look at utilization rates, look at the whole team, both humans and robots. Measure how much each worker contributes and assign work accordingly. Just as you wouldn’t hire a person to work only three hours a day, try to load as many tasks onto each robot as you can. The same workforce management and lean principles apply to digital labor as they do with people. Ultimately, if your robots effectively free up time in an employee’s day, you should put the saved time to use in higher value activities that robots can’t do. Alternatively, you should consider rebalancing the work and reducing headcount. Either way, your returns should benefit.
You’ll need guidelines. These days, you have to think about risk and controls from the beginning. (We discuss this in detail in our latest paper, “Who minds the robots? Financial services and the need to control RPA risks.”) Fortunately, digital labor tools can be connected to a ‘control room’ and they can be audited easily. But you still need to track and maintain what apps they touch, and to make sure robot IDs (and the people who control them) have appropriate access rights, and so on. Micro-RPA programs may use a lighter touch, but you still need to develop appropriate controls to maintain consistency and manage operational risks.
It’s time to rethink what RPA can do. Certainly, robots can be used to streamline and automate work typically sent offshore, and many firms are turning to digital labor to bring functions back in house. But as the tools become more sophisticated, they can now be used on a much broader range of activities. Micro-RPA opportunities generate small-time savings that can add up quickly. With digital labor, thinking small can lead to big returns.
Thank you to Charles Centrelli for contributing content to this viewpoint.
Partner, Intelligent Automation Leader, PwC US