5-minute read

Robotic process automation (RPA) — applying artificial intelligence (AI) and machine learning in automating routine tasks usually performed by humans — has emerged as the next leap in the enterprise’s journey towards digital transformation. According to Forrester Research, the RPA market is on pace to reach $2.9 billion by 2021, up from just $250 million in 2016. In a recent survey of operations decision makers, 74 percent reported having at least one RPA pilot underway with plans to roll it into production.

With an array of new, cost-effective tools now available, businesses are eager to transition repetitive, rule-based tasks to digital workers and to enjoy RPA’s promises of cost efficiencies, time savings, and quality improvements (e.g., reducing the risk of human error). However, as with traditional automation platforms, successful implementation of robotic automation requires choosing the right processes to automate, scrubbing and preparing your data, and ensuring continuous governance.

Identifying business processes for RPA

While robotic processing automation can be an ideal solution for organizations looking to lower costs and improve productivity, not all business processes are ideal candidates for handing over to digital workers. Most businesses find that RPA delivers the greatest benefits when used to automate processes that meet the following criteria:

  • Based in well-defined rules: Digital robots, like their mechanical counterparts, function best when they have a set of unambiguous rules to work with — the fewer exceptions, the better.
  • Access multiple systems: One of the biggest advantages RPA offers over both traditional automation and human effort is its ability to process inputs from multiple systems simultaneously.
  • Subject to human error: When a human worker performs a low-skill task over and over — like copying and pasting data — the risk for error can become dangerously high. Digital workers can perform the same routine tasks 24 hours a day, seven days a week, with a high level of accuracy.
  • Limited need for human intervention: Processes that a robot can execute from beginning to end, such as invoice processing, are ideal candidates for RPA.
  • Not subject to change: The investment of time and effort in programming robotic platforms is more likely to pay off if the process being automated is stable and not likely to change significantly.


Preparing data for robotic automation

Digital workers function on data, and quality data yields quality outputs. Organizations looking to leverage RPA will want to consider the following three initiatives before proceeding to the implementation process.

1. Consolidate duplicate records

Recently one of our clients requested our help in integrating their enterprise customer data from several platforms. As we reviewed their data, we discovered numerous instances in which new accounts had been created for different variations of the same company (for example, “Acme,” “Acme Company,” “Acme Widget Company, Inc.,” etc.) and for different locations. As a result, the client had approximately 200,000 enterprise customer records that involved some form of duplication.

Had this client chosen to implement an RPA platform using their data in its original state, not only would they have wasted time and processing power on thousands of duplicate records, but significant manual effort would be needed to review and correct the output, reducing or even eliminating the efficiencies the enterprise was hoping to achieve.

Looking at one of the most common use cases of RPA — invoice automation — duplicated data can result in an enterprise customer being double-billed for a single product or service. Or the customer may receive dozens of invoices for products purchased across the company, each requiring separate processing. By consolidating duplicate records before going down the RPA path, enterprises can ensure that their digital workers have a streamlined database to work with, resulting in greater efficiencies and a reduced risk of errors.

2. Eliminate orphaned data

Just about every organization has had the experience of a CMO or CIO initiating a massive data-collection project, only to have the initiative be canceled or abandoned. Even after the project goes away, the data usually remains stored away somewhere — unused, unexamined, and undocumented.

There are perils of the “retain everything” approach to data storage from a compliance perspective. The same orphaned data that can get an organization in trouble under privacy laws can also bog down RPA platforms with outdated information that no longer serves the business’ purposes. A comprehensive data audit offers the opportunity to shine a light on forgotten data stores in all areas of the organization, assess their usefulness (or lack thereof), and deal with them appropriately before incorporating an RPA platform into data processing operations.

3. Update inefficient processes

If a process is inefficient to start with, robotic automation will simply add another layer of technology — it will not fix the underlying problems. Before an organization spends the time and resources to implement an RPA platform, a thorough evaluation of the process will help identify technologies that need to be updated and procedures that can be streamlined. Once the organization eliminates these inefficiencies, implementation of an RPA platform can deliver on its promises of lower costs, faster turnaround, and improved quality.

Facing increased competition and escalating customer expectations, businesses across industries are under pressure to reduce costs, improve productivity, and improve quality. Robotic process automation has the potential to deliver on all three counts. When organizations lay a solid foundation by identifying the right processes, preparing their data, and, as we will discuss in a future Insight, ensuring continuous governance, they set themselves up to realize the benefits of RPA, both now and in the future.

Like what you see?

Paul Lee

Executive Team member Kevin Moos is recognized for his experience with knowledge management systems. He has lent his expertise to several prestigious industry panels on enterprise content management and other topics.