Robotic Process Automation: Reality Check and Route Forward
Do you have high expectations for RPA?
Early RPA initiatives have produced some tangible successes. Many companies, however, have yet to scale and exploit their RPA use to a level that delivers sustainable process improvement. Whether exploring RPA, launching a pilot or trying to maximize a robust solution already in place, it is vital that finance, HR, IT, procurement and global business services professionals understand what practices are critical to success. With this in mind, Canon Business Process Services is pleased to offer you an important research paper from The Hackett Group. The paper spotlights the current state of RPA adoption, key implementation factors that drive business results and how RPA fits within the smart automation model of the future.
It has been about five years since analysts began seriously exploring robotic process automation and its possible impact on business operations. This new class of automation technology brought with it the enticing potential to use software robots to carry out transactional, rule-based tasks on computer systems. Furthermore, RPA solutions offered the prospect of high returns on a very moderate initial investment. Many vendors claimed RPA could deliver process savings of as much as 60%-80% along with major improvements in quality, productivity and speed. Better yet, RPA could be deployed quickly with only limited IT organization involvement, if at all, for implementation or support.
At first, RPA projects targeted tasks in industry-specific processes, such as processing insurance claims or bank loans, before being embraced more broadly by traditional business services functions. Early adopters used it to automate tasks such as entering data into multiple systems or validating and verifying information such as vendor name, purchase orders, or duplicate invoices in accounts payable processing.
By now, many function leaders have seen evidence of RPA’s potential, but few have taken steps to exploit it as a means of sustainable performance improvement.
The State of RPA Today
Broad-based adoption of RPA is still nascent, but the number of initial success stories, both with large-scale adoptions and smaller, more targeted projects, is encouraging interest. Functional executives expect RPA use to increase significantly in the near term. While all functional areas studied foresee further use of RPA, finance and global business services executives are particularly bullish. In all, 78% of finance executives expect some deployment of the technology, whether limited (i.e., within pockets of the organization, or in specific processes) or broad-based, within two to three years, a sentiment echoed by 85% of GBS organizations.
There are several distinct RPA deployment models and use cases today. Scaled deployments are large, sustained initiatives with robots numbering anywhere from tens to hundreds, where they are a key part of service delivery and have a fundamental impact on the overall business. For example, bots are employed in account-to-report tasks such as balance-sheet reconciliations, consolidation and journal posting, management reporting and master data management.
Unlike scaled deployments, tactical deployments tend to have a medium-term time horizon of two to four years, but no imminent plans for replacing the bots. Beyond generating a financial ROI, the solution brings other benefits as well, such as higher process quality and easier auditability. For example, a healthcare provider implemented seven bots in its revenue-cycle management process to support an aggressive acquisition strategy. In the meantime it is pursuing a longer term goal to implement a comprehensive revenue-cycle management application.
Finally, disposable deployments are of limited duration and are focused on producing a payback before disposal, usually within six to 12 months. Examples include a data migration robot supporting a company merger, or a robot that administers staff-training records only temporarily while a new learning and management system is completed, eliminating the need for that task.
Ingredients For Success
RPA is most advantageous when applied to tasks with certain characteristics: they receive digital inputs; they use structured data; and the work to be automated follows unambiguous, logical rules rather than allowing discretion and judgment. Examples of such tasks include:
- Extracting and entering data
- Processing and updating forms
- Merging, consolidating and archiving
- Tracking, monitoring and archiving
- Formatting and reporting
- Downloading, updating and uploading files
- Conducting periodic analysis, performing calculations and preparing analytics reports
Approaching RPA Today Based on Lessons Learned
Business functions that have successfully deployed RPA beyond the pilot or trial stages share a number of characteristics. First, they are disciplined about identifying and assessing suitable RPA opportunities. Next, they plan the introduction and expansion of RPA based on a lifecycle perspective. Finally, they support their programs by building an RPA team or center of excellence.
Selecting suitable opportunities for RPA
An effective approach for identifying suitable tasks is to deconstruct a back-office process. In the example illustrated (Fig. 3), order-to-cash, purchase-to-pay and account-to-report processes are broken into several subprocesses. Specific activities within each subprocess will be analyzed and categorized as a strong, good, moderate or poor candidate for RPA. In the case of account-to-report, the example analysis finds that three subprocesses are strong candidates for using RPA because, at a high level, they involve extracting and entering data. This makes cash application, overall, a strong potential target for RPA.
It is important to note that Fig. 3 is illustrative, covering three high level process examples. The areas of analysis will vary based on company circumstances and can include back-office processes in any function, such has the recruit-to-retire process in human resources. Moreover, this represents just the highest level of filtering.
Lifecycle-based planning
A well-planned roadmap is essential to attaining the expected payback from RPA, first when it is introduced and later when it is scaled up. A roadmap ensures that, among other things, sufficient rigor is applied to areas that have proved challenging for others.
Examples are:
- Selecting the best tool for a specific use case.
- Building realistic expectations within the organization about expected ROI.
- Managing the politics of associated changes (e.g., in roles).
- Establishing sufficient resources to facilitate scaling, for example through training or a COE.
- Establishing governance for management of the bots throughout their lifecycle.
Building an RPA team or COE
As a business function gains more experience with and begins to extend its use of RPA, it needs a team of resources to support process owners who are introducing RPA into their own operations. The purpose is to leverage an existing body of knowledge and experience and thus avoid some of the issues that limited early adopters from realizing the full potential of RPA. In particular, it facilitates better assessment of opportunities, creation of repeatable automation capabilities, and coordination of multiple RPA projects across the enterprise.
An RPA COE can take different forms depending on circumstances, but should always be made up of a multidisciplinary team with leadership and political skills, business analysis and design skills, and technical development skills. Participants should also come from the IT function and relevant business function and process leads. Even though RPA deployment is usually business-led, support from the IT function is vital to enabling and scaling RPA by providing guidance, requirements and resources for infrastructure, connectivity and operations management.
What Comes Next? The Evolution of RPA and Smart Automation
As RPA technology matures and starts to offer more functionality, and as organizations develop related skills, knowledge and experience, adoption will rise. However, even at this early stage, The Hackett Group’s research shows that many organizations are beginning to approach RPA as part of a broader enterprise digital strategy. A critical element of this strategy is the transformation of operations based on the principles of smart automation – faster, incremental, business-driven – using a combination of emerging technologies that includes RPA.
Smart automation requires a solid foundation of enterprise automation, workflow, RPA and cognitive technologies. While most organizations have mature capabilities for automating common transactional processes and standard workflows, generally speaking the business functions remain in the very early stages of adopting cognitive or artificial intelligence technology; almost all that currently use these capabilities do so on a limited basis. According to our research, adoption rates will grow significantly within two to three years. Many organizations are now piloting cognitive computing and artificial intelligence projects, including 47% of GBS organizations and 38% of finance organizations.
Looking Ahead
While scaling up and expanding adoption of RPA has been more challenging than expected, this technology is still poised to play a key role in digital transformation. However, RPA will be just one of a number of tools used to automate and integrate transactional, decisioning and knowledge work. To plot a path ahead and achieve step-change performance improvement, business functions must begin thinking more broadly and longer term, developing a vision and roadmap for smart automation that includes both strategic and tactical roles for RPA.
At the same time, a reality check is warranted. A certain amount of trial and error is necessary to learn the capabilities and limitations of these tools, but companies that expect they will be in broad use in the next few years—as our research suggests many do—must exercise discipline when analyzing opportunities and developing business cases as part of a robust process for managing the RPA lifecycle.