BY: Ben Codoroli
With use cases in virtually every industry and department, RPA has quickly become a popular tool for companies to revolutionize the way they think about their business processes. Global research firm Gartner has named RPA the fastest growing segment of the global enterprise software market. RPA revenue grew 63% in 2018, and Gartner expected that 2019 would be another huge year of 54% growth to $1.3B.
So why, according to a study by HfS2, is RPA satisfaction among Global 2000 leaders sitting at 56%? For such a hot segment, shouldn’t this be higher? Rather than trying to bury a statistic like this by shifting the focus to the staggering industry growth or the mind-boggling ROI numbers, I think it’s critically important to the future success of RPA that we dissect what might be going wrong and how this can be avoided.
The first culprit for the mediocre satisfaction ratings lies with the overzealous marketing pitches in the RPA industry. One of the more common stories presented to potential customers, who innocently soak it up with stars in their eyes, is that “bots that were built in hours are now saving clients thousands of man-hours per year.” And while there certainly are legitimate stories like this out there, this doesn’t just happen overnight. There must be infrastructure in place to support an automation program that generates and maintains such results. What happens if the bot breaks? Or if the process changes? How was this process that was such a perfect candidate for automation discovered and chosen in the first place?
Don’t forget that each new application or website you want RPA to interact with introduces new uncertainties to how seamlessly the RPA software will work with it. Selectors may be unstable; buttons and text fields may be embedded objects that are poorly labeled in the underlying code. These issues don’t mean that RPA cannot be used for the application—it just means it might take longer than initially expected. Plan for this uncertainty and don’t promise lightning-fast turnaround if you don’t have experience using RPA with the application in question.
And lastly, know the strengths and limitations of RPA vs other automation tools. Just because you have an RPA team does not mean that every process needs to be automated using RPA. Sometimes companies get so excited for RPA that they say “yes” to every process that gets pitched to them, even if there are better ways to do it. For instance, I’ve seen some cases where a simple Excel macro or R script could have automated an entire process, but the RPA team felt like they had to use their RPA software just because they could. The result was a project lasting weeks rather than days, and left some people scratching their heads as to why things seemed to be overcomplicated.
In order for RPA to be taken seriously and to continue on its strong trajectory, we must be honest about it and set realistic expectations. It is not some silver bullet that can be implemented in one week and then solve all of your process woes.
Siloed Automation Teams
Let’s say a company completes a successful pilot phase implementing RPA to automate a couple of processes, and they green light efforts to expand and continue to automate more processes. A big mistake is made if the next steps do not include interdepartmental communication and partnering.
An internal marketing plan should be used to educate employees on the capabilities of RPA to include everyone in this new wave forward. This is important for two reasons. The first being that, despite the absence of any meaningful evidence, there is a concern around RPA (along with automation as a whole) that it is “coming for your job.” In a global RPA study performed by Protiviti3 in 2019, 450 companies that use RPA in various capacities were asked about the top benefits of RPA use. “Reduced costs” received a measly score of 1% among those considered RPA Leaders. Compare that with “Greater employee satisfaction from elimination of mundane tasks” receiving a score of 12% among the same group. The fact is, RPA implementation should make all employees excited. In the vast majority of use cases, RPA is taking away the most boring, repetitive tasks of an employee’s day off their plate. The key is communicating this with your people so they can envision the more interesting tasks they’ll be able to focus more of their time on rather than fearing for their jobs. The second reason for this company-wide communication is to generate new ideas for automation. Generally, the best ideas for automation will come from the “boots on the ground”—those who actually know how much time they’re spending on frequent, repetitive, rule-based tasks. Educate them on RPA’s capabilities and get them excited to share their ideas that will fuel the RPA movement.
An RPA team that does not have ardent, visible support from higher management will find it difficult to get things done. Think about what is needed from an IT perspective for an RPA program to scale: servers hosting the orchestration components, virtual machines to host unattended bots, service accounts to systems that the bots will use when performing certain processes, etc. Now we all love our IT folks—but they don’t always love it when we want to make a bunch of changes to the infrastructure, provision new resources, grant new account permissions, and so forth. Pair your requests with the fact that there’s a decent chance they have not encountered RPA before, and you’ll find yourself explaining the entire concept to five different people for each request that you make. This will not only grind any project to a painfully slow pace, but it will also start to create tensions with the team that RPA may critically rely on periodically throughout its life cycle. By setting the tone from the top that RPA is an important initiative that will require collaboration from various departments, these frustrating interactions can be exchanged for clear and productive conversations.
Companies also fall short of RPA satisfaction when they focus solely on automating existing processes without consideration for process improvement activities. In fact, process improvement should not only be a consideration when implementing RPA, it should be engrained into the mindset of the entire team. For real change to be made across the organization, a wholistic approach must be used to eliminate wasteful steps and to optimize and standardize remaining steps where possible. This increases the probability of a successful bot deployment by consolidating steps—resulting in a quicker build—and lowering variability of the process resulting in quicker development/testing and a lower chance for errors or exceptions during the bot’s routine execution of the task. Focus on the value of each process step. Don’t blindly automate every process that comes your way.
One of the more popular use cases for RPA is to connect older systems that can’t otherwise communicate with each other. This typically results in quick wins, solving problems that have caused much frustration for employees that were stuck having to act as the bridge themselves. These processes are great to automate and typically make the best pilot processes because they are relatively simple to build and the results can be easily measured. However, companies would be mistaken to focus all of their efforts on these types of processes without also setting their eyes towards bigger goals—AI.
As the RPA industry grows, it’s apparent that the big players are heavily investing in expanding their built-in AI tools. Automation Anywhere has IQ Bot, and UiPath recently rolled out Document Understanding. These tools leverage AI to allow customers to feed in documents such as invoices or receipts that may not be of the same format, and the robot is able to parse out the important fields. Then as a user validates the data being extracted, the robot is trained to become more and more accurate over time.
Experts estimate that 40% – 80% of all data in any given company is in an unstructured format. This includes text files, images, emails, etc.—data that cannot be easily aggregated using traditional methods. Machine learning and AI are the tools used to unlock value from unstructured data, but the time, effort, and skills required to build such tools have traditionally been a hurdle for companies to take advantage of these technologies. RPA is beginning to change that by making AI/ML tools more accessible to wider audiences.
So yes, continue to automate the more traditional processes, but don’t neglect to take advantage of the AI/ML capabilities as well. Harnessing these technologies will help your RPA program provide more meaningful and lasting impacts across your business.
By avoiding common pitfalls, RPA can achieve great success
The impacts are real: companies really are removing the mundane, repetitive tasks from their employees and shifting their time to more strategic or client-facing initiatives. But we can’t talk only about the success stories without being realistic about the investments needed to truly realize RPA’s value. RPA is an effective tool that, when paired with proper governance, systems, and policies, can lead to greater efficiencies, less errors, and higher employee satisfaction.
1Gartner Says Worldwide Robotic Process Automation Software Market Grew 63% in 2018. https://www.gartner.com/en/newsroom/press-releases/2019-06-24-gartner-says-worldwide-robotic-process-automation-sof
2HfS study of 355 Global 2000 operations leaders conducted with the support of KPMG: “State of Operations an Outsourcing” 2019. https://www.horsesforsources.com/
32019 Global RPA Survey Results. https://www.protiviti.com/NL-en/insights/rpa-survey