Robotics Process Automation has arrived. But what does this mean for organizations? Row after row of shiny androids performing complex functions? Not quite. Instead, the proliferation of robotic process automation (RPA) is the more likely scenario, empowered by shared services centers (SSCs).

Automation and AI technologies have made significant in-roads across business sectors. Innovators are discovering new ways to automate routine tasks such as data entry and reporting, while artificial intelligence initiatives allow AI to emulate portions of the human decision-making process. RPAs, meanwhile, are a fusion of these technologies; small, semi-autonomous processes capable of performing specific, high-level tasks.

According to a recent ADP/ScottMadden webinar, Robotics Process Automation: Practical Tips for Implementing RPA in Your Organization, RPA is being used in front line operational processes, but is largely gravitating to back office processing in SSCs as a way to cut overall costs, reduce error rates and improve efficiency. Why SSC's? These models vary by company, but the common denominator is centralized volumes of transactional activity — the "sweet spot" for exceptional RPA value. Although there's a good reason for the hype surrounding this technology, you shouldn't just jump into the fray without a solid game plan.

1. Start Small

The first step in RPA integration is moving a robot (or two) into production for four or five back office process scenarios. The move to production does not require an overhaul of systems, but will require some IT oversight and testing support. This is a relatively new concept, so a "pilot" will deliver better and faster answers as they apply to you than surfing the web. While your results may vary based on business objectives, cultures and processes, you're looking to answer some basic questions, including:

  • What types of licenses are required? Many process vendors offer cheap or free "trial licenses" for their RPAs to prove out your pilot. Production (purchased) licenses often include attended bots you kickoff, unattended bots that kickoff based on programmed signals, coordination bots that manage those signals, and others.
  • How much RPA scripting fits on a single license? Once you ramp up production, clarify with vendors how much scripting fits on each license; do you require a licenses for a single process? Or can you fill a license up with numerous processes?
  • What do setup and integration timelines look like? Do you require days, weeks or months to get a single process up and running? What attributes of processes take longer than others?
  • Can robots run on laptops or are servers required? If laptops, what speed and complexity limits exist? How much will server space cost? What are the IT protocols to set up these environments?
  • What processes offer the best return on investment? Researching and forecasting the answer to this question before you get started can be mission critical to attracting additional stakeholder funding, or suffering a "ho-hum" reaction.

2. Find Real Issues

The goal here is to identify a real problem in the organization that RPAs can help reduce or eliminate. Find an issue that's repeatable, regularly occurring and comes with significant "volume." As the webinar points out, it's better to start with "rivers" than "creeks" — robots do well "buzzing through" high-volume, logic-driven tasks.

First, find a champion for the RPA cause. Often, immediate managers and supervisors struggle to find time, while C-suite members are too removed from the work itself. Your best bet is likely to find a director or vice president with the knowledge and influence to provide ongoing support.

Next, work with IT to build out a test bed and discover how your RPA performs — and make sure you hit "save" when robots succeed — this "scripting" can be re-used to expand to other sub-processes. Demonstrating the support needed from IT, the time it takes to build and test and the practical value of these processes can help silence skeptics and overcome organizational resistance.

3. Get Ready for Some Resistance

Now it's time to go live and to move the RPA you've worked so hard on into the production environment. This is where the real learning begins. It's one thing to secure support from process owners, IT, continuous improvement groups, etc. in a "test" environment, but it's another to ask for ongoing accountability when the robots go into production. Expect some conflict and/or confusion over who "owns" the robot, who maintains the robots, which processes are next in line for automation, who can develop robots, who monitors licenses and many other practical governance roles that must be ironed out in a production environment, but not necessarily in a test environment. The answer with most is some hybrid of centralized authority (typically license purchase, testing/quality assurance, robot coordination), and decentralization (process selection, robot development). Too much centralization leads to bottlenecks and dissatisfied functional owners. Too much decentralization can lead to robot chaos.

4. Prepare an Enterprise RPA Program

Prepare your organization for the adoption of RPAs by surveying and quantifying transactional work volumes, then look for high-volume, high-impact, mid-risk processes that are a good fit for robots and build out the business case based on pilot results. Make sure robot governance is defined, data is relatively clean and rollouts won't negatively impact big decisions already in the pipeline, such as changes to business process outsourcing or system plans.

5. Develop an Overarching Digital Strategy

RPA is not the only "disruptive technology" out there. We are seeing more use in our everyday lives of virtual agents that can converse with humans by phone and tap into systems or databases to provide answers. We are also monitoring the use of Artificial Intelligence capability to recognize complex patterns, and pull vast quantities of unstructured data to help us make better decisions for higher value processes — think financial forecasting, market assessments, credit risk, workforce planning, benefit design, supplier spend analysis, etc.

All of this warrants a complete re-look at the way you are designing your support and operational services to be delivered across the enterprise. Factor in the pace this technology is moving. Five years ago, it was hard to find a process robot, today they have been implemented in about 60% of our organizations. The robotics revolution is here and now and it only makes sense that robotics and automation strategies are represented in your future technology plans.

About ScottMadden's Corporate & Shared Services Practice

ScottMadden, a general management consulting firm, has been a pioneer in corporate and shared services since the practice began decades ago. The firm's clients span a variety of industries from entertainment to energy to high tech. Examples of ScottMadden's projects include business case development, shared services design and shared services build support and implementation. ScottMadden has partnered with ADP, a leading provider of human resources management software and services, to help educate mid-to-large organizations on implementing robotic process automation (RPA).

Tags: Automation onboarding