How to Avoid the Pitfalls of Automation
Robotic process automation is typically used to streamline mundane and repetitive business processes to gain efficiency and free up workers' time for more meaningful tasks. Tasks that can be accomplished using RPA run the gamut, from data entry and tracking orders to sending emails and making phone calls, to filing paperwork and sequencing basic mouse and keyboard operations.
The use cases keep expanding. RPA usage is projected to grow significantly in the next few years, with the market expected to expand at double digit rates through 2024, according to a Gartner report.
The Benefits of RPA in the Workforce
The talent shortage may be an impetus for this growth as some companies look at RPA as a solution to worker shortages and supply chain challenges, said Melanie Nuce, senior vice president, innovation and partnerships at GS1 US, a barcode provider.
“With the unprecedented labor shortage acting as a main ingredient for ongoing supply chain disruption, RPA can help retailers offset the lack of human workforce and improve efficiency,” she said.
RPA can boost productivity, but it can also enhance the employee experience, according to James Matcher, intelligent automation leader for the Americas at global consultancy EY.
“One of the biggest learnings over the years is that organizations need to look beyond productivity and efficiency as a measure of impact and value from RPA,” Matcher said. “While these are certainly key areas of impact and value, there are many processes in organizations where RPA could also improve employee satisfaction and retention, increase agility and responsiveness in an organization, create an improved customer experience, accelerate revenue, reduce organizational risk or strengthen compliance oversight.”
There are numerous benefits to the technology, but those also come with significant risks.
Related Article: RPA and How It's Adding Value in the Workplace
Assess the Processes You Want to Automate
Bill Gates has been quoted as saying: “The first rule of any technology used in a business is that automation applied to an efficient operation will magnify the efficiency. The second is that automation applied to an inefficient operation will magnify the inefficiency.”
That's a sentiment Preeti Lobo, practice director, business integration and automation at Apps Associates, agrees with. Issues tend to arise when companies do not thoroughly evaluate the processes they wish to automate.
"Organizations that implement RPA to automate inefficient processes will not realize the benefits of automation but, instead, further propagate inefficiencies,” Lobo said. “It is important for organizations to evaluate the processes/tasks that they intend to automate first and bounce them off a few rules."
Among those, she lists:
- Is the existing manual process efficient?
- Has the process been performed manually for at least 6 months?
- Does the volume/frequency of the task make it a worthy case for automation?
- Is the task/process currently cumbersome?
- Does the process cause customer dissatisfaction or delays due to it being handled manually?
To avoid the pitfalls, Ron Cameron, CEO at St. Louis, Mo.-based software company KnowledgeLake, said organizations that are serious about RPA and looking for an outcome, rather than just wanting to try out the latest technology, should carefully assess their solution provider to ensure it is focused on the desired outcomes of the project. He suggested that companies:
- Identify a specific project that has clear goals before engaging a solution provider.
- Determine whether the RPA solution provider is focusing their marketing efforts on opportunities and outcomes, or simply selling their technology.
- Ensure the solution provider uses tangible metrics to measure success. Every RPA project should have a clearly defined set of metrics to define success.
- Ensure the RPA vendor will provide continued support once the solution is up and running.
- Commit to learning as much as they can in order to be knowledgeable about the vendor mix necessary to be successful with RPA.
The success of RPA depends on how serious the company is about its use and outcome. A commitment to learning not just about what RPA can accomplish today but how the technology — and the chosen vendor solution — will evolve is critical to the ROI in this market.
“RPA integrations for data exchange are quite useful, but it is downstream in processes where the real benefits occur — seamlessly driving data into the right systems and processes, and automating whole processes from end to end, as opposed to simply digitizing individual manual tasks,” said Cameron.
Related Article: 3 Ways Human-Machine Collaboration Increases Employee Productivity
Understand the Interplay Between AI and Automation
RPA is well-suited to mundane, routine tasks. More complicated tasks that require reasoning and decision-making, however, are exceedingly difficult to automate. Antti Karjalainen, CEO and co-founder of Robocorp, said the first versions of RPA accomplished much but their limitations soon became apparent.
“The first generation of RPA emerged in the early 2000s and immediately made an impact on businesses," he said. "However, as the world around Gen1 RPA evolved, it stayed stagnant in its capacities. Today, Gen1 RPA’s shortcomings have become more evident, as it is hampered by outdated payment models, the broken bot syndrome, vendor lock-in and an overall limited set of capabilities.”
Artificial intelligence, cloud services and open-source programming languages are now being used to create RPA bots that are able to handle much greater complexity. “With these problems recognized, we are now entering the next generation (Gen2) of RPA, where providers are seeking to mend these imperfections," Karjalainen said.
RPA is still limited to the specific operations it was programmed to perform. But the use of AI and machine learning in conjunction with RPA are beginning to open up the doors for innovation, although RPA is not able to truly innovate on its own.
“Operations requiring judgment or probabilistic decisioning are not well suited for RPA. The technology is not designed for that — however, those operations are typically preceded or include steps/activities that are well suited for RPA,” said Matcher.
Related Article: Intelligent Process Automation Pushes the Boundaries of Business Process Automation
Consider the Cost of Technical Debt
RPA is often thought of as a short-term gain because it requires constant updating as things evolve, from both a technology and process standpoint. RPA has come a long way since the shell scripts of MS-DOS and UNIX, but there are still many RPA bots today that manipulate a front end or graphical user interface.
Learning Opportunities
The ease of use that GUI interfaces provide enables practically anyone to create script-based RPA bots to perform a task — which is great until something on the backend or frontend changes or is updated, and the RPA ceases to function. This is what's referred to as "technical debt" and can end up adding cost through additional resources to get things back up and running. By creating RPA bots that are resilient to the changes that are inevitably coming, companies can avoid the risk of things coming to a sudden halt.
“Some organizations don’t consider the support of the tool once implemented," Lobo said. "Contingency plans are imperative. Evaluating long-term support models will help mitigate minor changes, restarts, and so forth, of the bots."
Matcher said RPA can be a long-term solution despite the fact that most businesses regularly update their processes. “There may be a cyclical processing of automation needs, but certainly no elimination of the technology altogether. The need will never go away,” he said.
In his view, automation stands out in the case for flexibility. “Where a business’s ERP system can’t keep up because it can’t be updated every three months, automation can handle updates every week if needed," he said. "Automation is your digital last mile to give you flexibility in your processes.”
Related Article: Why Enterprise AI Needs Human Intervention
The Impact on Jobs
RPA is extremely useful for accomplishing high volume, routine tasks. It can outperform humans and is more cost-effective. So, what does that mean for workers in the future? According to a McKinsey report, by 2030 up to 800 million jobs may have been replaced by automation.
“Rather than having an employee spend hours randomly scouring aisles of a store trying to handpick expired items off the shelf, RPA software can ingest, digest and disseminate data, informing staff that certain items need to be pulled,” said Nuce.
Better yet, a physical robot may be what's removing items from shelves, foregoing the employee entirely. "The time and financial benefits are palpable, as each retail operation can maximize the output from its workforce and free up employees from time-consuming and manual tasks" she said.
But there's no need for workers to panic just yet. Only 5% of occupations can be fully replaced and the majority of employees whose jobs will become automated are likely to be transitioned to a different role or trained for other tasks, according to McKinsey's analysis.
Although automation can replace a lot of processes that are part of an employee’s job, it cannot replace every part of any one job, according to EY's Matcher. “It has become very clear that it is very difficult to automate 100% of a person’s role and [it] is more likely to enhance and work alongside an employee, increasing their output, which for an organization is far more valuable than the cost savings by removing the employee," he said.
In other words, automation supports employees, rather than replaces them. That's because a human must still be responsible for monitoring the quality and accuracy of the work. “The increased use of automation and technology does not render humans obsolete," said Nuce. "Guidance at a human level is still imperative to ensure the outcome is achieved successfully," she said.
And not all processes should be automated from end to end. Apps Associates' Lobo suggested that processes should sometimes have human intervention to ensure that errors don’t process further and can be nipped at appropriate places, such as having a person review journal entries uploaded by a bot before they are approved.
Ensuring that people in the organization are educated on the benefit of the tools and how it can positively impact their own work-life and the company is also important, Lobo said. "Change resistance and/or fear of change could cause employees to not disseminate all aspects of the process, rendering the automation to be error-prone," she said.
Related Article: AI Is Not Coming for Jobs Just Yet
Though RPA has limitations and is not currently capable of true innovation, organizational leaders are finding they are able to use it to accomplish routine, mundane tasks and gain greater efficiencies. Clearly examining, and re-examining, the risks and pitfalls of automation is critical to get it right. And for leaders, it's important to be clear how change can be a benefit for workers.
Increased use of RPA can mean an opportunity for upskilling and reskilling with a focus on more interesting tasks. For the company, that leads to not just a boost in productivity and a reduction in errors, but an overall greater employee experience as well.