Key Strategies For Implementing Robotic Process Automation

New technologies such as robotic process automation offer the opportunity to automate repetitive processes and more efficiently achieve business goals. Deploying software robots can enable organizations to execute current processes faster and more accurately, 24/7. The benefits can be significant and include boosting the return on investment; improving operational accuracy; enhancing compliance; achieving cost reductions and increasing scalability.

RPA is changing many business scenarios, such as when a new employee joins a company. Typically in this situation, IT managers need to update the employee information in numerous systems, entering and copying the data from one platform to another. Or, consider managers in Human Resources who create candidate offer letters and onboard new employees following structured processes. These are examples of clerical tasks that employees perform every day—tasks that are mainly routine but demand a significant amount of time.

Robotic process automation is enabling organizations to release workers from these types of time-intensive, manual processes in order to focus their talents on more strategic activities that can drive the business forward. The key question is, could an RPA strategy be applicable to your company? If the answer is yes, what are some key implementation steps to consider?

Beware of The Challenges

In our experience at Canon Business Process Services, a good place to start is to be aware of some important challenges that you are likely to face when deploying an RPA solution. These challenges include the following:

Employee concerns. Upon hearing that process improvement or automation projects are on the horizon, employees may become concerned that their jobs will be eliminated. They are asking themselves the question, will you still need me? A structured change management program with well-crafted communication plans can address this concern.

Unrealistic expectations. This is often based on the initial excitement over new technology, especially RPA. This unbridled enthusiasm can cause business leaders to have the impractical belief that just about everything can be successfully automated; that RPA might represent a silver bullet. To counter this thinking, a best practice is to be conservative when clarifying success criteria and realistic in defining the business outcome to senior leadership.

Lack of skills. Does your company have the internal expertise to deliver this new and emerging technology? The likely answer is no. This includes the ability to deal with the daunting and frequently confusing task of selecting the best RPA solution for your needs from among the many alternatives available in today’s market. However, the good news is that there is a wide range of RPA tools and service providers that you can leverage to help solve these and other issues.

Leveraging IT. Do you implement RPA on premise or do you consider a cloud/hybrid-based solution? Either way, your RPA project will most likely need support from your IT and data security teams. It is a common misconception that the IT department won’t be needed. On the contrary, whether the company is implementing RPA on premise or off-site, some internal infrastructure and/or security issues will need to be resolved.

Dirty data. Redundant, outdated and trivial data (commonly referred to as ROT) has always been a challenge for enterprises. This is especially the case, however, for software robots (or BOTs), which need accurate data in order to efficiently follow the business process parameters that have been set. If data is missing or inaccurate, a BOT will not be able to correct, update or complete a process, which creates an “exception” requiring time and resources to resolve.

Key Strategies For Addressing The Challenges

The next step is to examine key strategies for meeting the challenges spotlighted above. Let’s begin with employee concerns about job security. This fear is certainly understandable when you consider the business benefits that BOTs offer.

Mitigating Employee Concerns

For example, there’s speed: a BOT can complete a business process task on average 10 times faster than a human. Another factor is time: BOTs can run 24/7/365 versus the average eight-hour, five-days-per-week schedule of most employees. Next is expense: the cost of a worker to perform a job including benefits is much higher than a BOT, even if offshoring is considered. Finally, let’s look at accuracy: workers are human and make mistakes, particularly when repetitive tasks are involved. BOTs can repeat the same task exactly the same way millions of times with no error.

With such a strong case for BOTs, how do business leaders mitigate employee concerns about how a new RPA project may impact their job? One answer is to communicate the fact that deploying RPA and software robots can enable employees and the overall enterprise to operate smarter and more efficiently and to devote more time to activities that have greater strategic value to the organization. This can translate into a workday marked by reduced errors, with automation performing repetitive tasks and documents or data flowing to designated locations without any manual touch points in between. That is a scenario for using technology to better ensure the success of the company and its employees going forward.

Setting Realistic Expectations

As noted earlier, it’s relatively easy for some executives to think of RPA as a silver bullet that can be implemented to automate virtually every business process in the organization. Such a viewpoint could result from the growing awareness that RPA has helped companies substantially reduce head count, slash average business process handling times and dramatically improve customer satisfaction. While there are some large scale, big budget success stories—and more occurring every day—a better approach is to set modest, realistic expectations with executive leadership, sponsors and stakeholders about anticipated results for your first RPA project.

This means clearly defining success criteria. For example, RPA project leaders often promise to reduce head count. This rarely happens. The more likely scenario is that instead of reducing head count, working hours will be diverted to higher-value activities. This expectation is related to the issue of employee job security. The company can realistically communicate that a targeted number of working hours will re-focused on more strategic activities—rather than that workers will be replaced by BOTs.

However, setting realistic expectations can go beyond promoting the idea that RPA can save time. There are other measurements that can be part of your success criteria. These may include a reduction in processing errors; the opportunity for knowledge workers to better focus on higher level business functions and enhanced customer service and faster processing cycle times.

Whatever results you target, best practices include being crystal clear about what the RPA project is designed to achieve and documenting these expectations in a project charter. This will enable you to refer back to the success criteria once the project has gone live.

Leveraging the Right Software and Skillsets

There are many RPA software providers in the market, each offering areas of special focus. This makes trying to select the right vendor when starting your RPA journey a daunting task. Also, the RPA software market and the skills required to implement these offerings create a complex landscape that is far from static. New players constantly appear and acquisitions are common. There is certainly no “one size fits all” approach available for an organization launching an RPA pilot or proof-of-concept project.

In our experience a successful RPA pilot program takes into account three important elements: people, process and technology. These factors are vital when selecting a managed services partner to help kick-start your RPA journey. Because all three components are equally important, it would be a mistake to think about your first RPA project only in terms of technology. Simply teaming up with a vendor to add technology to an existing process yields a good chance of inflating current process costs without any significant gains.

In addition to technology expertise, it is just as important that you work with a services provider that has experienced people who can help you identify the right processes likely to generate initial success and quick wins. One way to think about leveraging the right software and skillsets is to visualize your strategy as a three-legged stool. One leg represents people, one signifies process and the third implies technology. Without all three legs (or strategy elements) the stool falls over, which means there is a great likelihood that your RPA project will not achieve the intended results.

Including Your IT Department

When it comes to implementing RPA, solving infrastructure and security issues is critical. The best practice here is, again, correctly setting expectations. There are some industry articles, brochures, briefs and marketing materials on the subject of RPA that suggest that your IT department will not be needed for implementation.

Instead of proceeding with this assumption, a safer approach is to reach out to your IT and security teams for their input. At Canon we have implemented a number of hybrid approaches to help reduce the level of internal IT involvement. However, in our experience, even with this approach a company will need some amount of internal IT support in order to help ensure a successful deployment.

Obtaining Accurate Data

Many older and larger companies have to contend with issues caused by dirty data. While ensuring clean data can be a challenge for any technology project, it is particularly the case for RPA implementations. The reason is that BOTs are not as smart as people when it comes to diagnosing and solving problems due to incorrect information.

Another complication is that dirty data take on a variety of forms, some of which include the following:

• Missing data: including mandatory fields that have not been filled in or data that has not been validated (two likely scenarios when legacy systems are involved)
• Inaccurate data: incorrectly entered data (typically caused by lack of quality control)
• Inconsistent data: the same data that has been entered differently (sometimes caused by multiple owners of the data)
• Duplicate data: multiple records of the same data (a situation that can also be caused by multiple owners of the data)
• Orphaned data: data missing a parent record or parent data missing child data

How do you address the problem of dirty data in order to ensure a successful RPA pilot project? There is no easy answer when it comes to correcting, updating or removing duplicate records. A significant effort is often necessary to remediate legacy data. If you are planning a large-scale, long-term RPA project, it may make sense to launch a data cleansing project first, particularly if you are aware that your organization has data issues. At Canon we use a combination of automated and manual approaches in concert with our Business Processing Centers to remediate data for our clients. Based on this experience, following are two approaches we’ve utilized to help clients tackle the dirty data problem when implementing an RPA pilot program or proof of concept.

The first strategy is one we applied when implementing RPA for a large financial company that processes workers’ compensation claims. The project included training BOTs to look up data in a legacy system that contained a list of employers. The data in the system, however, was far from accurate, containing duplicate and inaccurate employer names such as ABC, ABC Corporation and ABC Inc. Consequently the BOTs were not consistently matching employer names in claims forms with the names listed in the legacy system.

The solution was to enable the BOTs with fuzzy searching capability and over time continually “tune” the level of fuzziness to the point where the “hit rates” improved to 70% or greater. This solution yielded a fairly quick win that increased the client’s data matches and ultimately reduced the amount of manual review required by the Canon project team. So the key tip is to experiment with fuzzy search methods to ensure that BOTs are consistently and accurately matching important data.

A second strategy is to consider using a subset of data that you know is clean and appropriate for your pilot project. In summary, the key takeaway for ensuring accurate data is: be prepared to review your data prior to implementing RPA.

New and innovative approaches like RPA are all about re-imagining your business processes so your enterprise is better prepared to thrive in the days ahead. Whether your RPA initiative is managed internally or in partnership with a managed services provider, the best strategy is one that is designed to automate your workflows and achieve significant, measurable performance improvements. At Canon we combine human ingenuity with the power of today’s most advanced software BOTs, enabling the two to work together intelligently. The result: by automating routine business processes, your employees can devote more time to serving customers and other higher-value work. That’s a powerful scenario for being future ready.