Robotic process automation, or RPA for short, burst onto the enterprise tech landscape a handful of years ago, promising to replace humans with robots or “bots” – scripts that automate human interactions with a user interface or application programming interface.
In spite of their futuristic name, however, bots are in fact little more than simple scripts, similar to the macros familiar to Excel users or Macintosh aficionados as far back as the 1980s.
Bots, in fact, have two fundamental shortcomings: They’re brittle, and they do nothing to lower organizations’ technical debt – and can even add to it.
Nevertheless, the RPA market has exploded. The unicorn among unicorns, UiPath Inc., went public earlier this year, and continues to experience dramatic growth. Clearly, the company is doing something right, in spite of the limitations of RPA.
To get to the bottom of this conundrum, I headed to Las Vegas to UiPath’s Forward IV conference, keynoted by Chief Executive Daniel Dines (pictured), to see what the fuss was about. Here’s what I learned:
Why customers love UiPath
The primary business benefit from RPA is improved efficiency. Employees can leverage bots to automate otherwise routine, repetitive tasks, allowing individuals to spend their time on more valuable, interesting work.
Calculating the ROI from such efficiency gains is straightforward and typically more than covers the costs associated with implementing and running UiPath’s platform. Add morale and retention benefits to the mix, and RPA becomes a no-brainer.
There are other RPA vendors in the market, but UiPath has rushed to the head of the pack. In part, we can attribute this success to the quality of its bots. Scripting interactions with the wide variety of user interfaces in enterprises today is a hard problem to solve that gives UiPath a barrier to entry in the marketplace.
Speaking with several UiPath customers, however, clarified the primary reasons UiPath has been so successful: its enterprise-class sales and support, as well as the support of its systems integrator and vendor partners.
Today, enterprises that have moved beyond small-scale RPA experiments to considering broader deployments have learned first-hand that the technology is brittle and has technical debt issues. UiPath customers are under no illusions about the fact that bots can break if something changes.
The solution: comprehensive, full lifecycle support and enterprise-grade governance, respectively – both from external sources as well as the information technology organization. “Engage IT,” advised Joe Bohnert, Intelligent Automation COE manager at Facebook Inc.. “At the end of the day this is technology.”
The support generally begins with UiPath and its SI partners, but in many cases, it includes the buildout of entirely new RPA divisions within IT organizations. UiPath’s RPA requires a team of developers, as well as every other role necessary for building and running enterprise software, including architects, testers, help desk and operations personnel, and more. In spite of this necessary investment, customers report solid return on investment – as long as they put the proper support and development resources in place.
The limitations of UiPath’s RPA
The most significant limitation of UiPath’s RPA is that in spite of its name, it’s really more about task automation than process automation. “RPA has to happen at the task level,” explained Junaid Ahmed, corporate vice president of finance at Applied Materials Inc.. “You can’t automate everything. You need business processes to handle that.”
Ambiguity over this terminology is partly to blame – after all, how many tasks do you need to string together before you have a process? Nevertheless, with UiPath, automations focus on individual tasks that one person might undertake.
For example, a common automation is to take data out of one application, perhaps perform a calculation and then input it into another application. From the perspective of the human, these three steps make up a task, and a bot is certainly able to complete it.
In contrast, end-to-end workflows and other enterprise processes are not the focus of RPA – at least, for now. Although UiPath claims it is not looking to compete in the business process management market, it is adding long-running tasks (which you might call processes) to its repertoire, and is certainly talking about enterprise process automation.
Other challenges for UiPath concern perception more so than reality. In spite of robots’ sci-fi cachet, many people are uncomfortable with the term – especially at work. Perhaps sticking to a term such as “script” or “macro” would have eased such fears.
There is also a widespread belief that automation will put people out of their jobs. Many conversations and presentations at Forward IV focused on dispelling this myth.
In reality, automation changes how people work by relieving them of time-consuming, repetitive chores. However, there is no reason to believe employers are firing people as a result. Instead, employees get to focus on more valuable and interesting tasks, increasing their worth to their companies and improving morale.
Perhaps the greatest challenge facing UiPath is the maturity of its product line. Its growth has outpaced its product development, giving it what basically amounts to growing pains. As a result, it is actively pursuing category expansion, both via acquisitions and organic growth.
The future of RPA
Today bots can automate tasks, but humans generally must create the bots manually.
There are shortcuts, to be sure — for example, a developer can record the actions of users as they navigate various screens. UiPath’s low-code bot creation capabilities are also an essential enabler that speeds up and simplifies bot creation.
What organizations really want, however, is an automated way to create the automations – a goal that won’t become a reality until the science of artificial intelligence matures.
Ideally, a business user would only have to express the intent of the automation, and the AI-based automation routine would build it automatically.
UiPath does have rudimentary AI capabilities, for example text-based sentiment analysis and entity extraction. Today’s AI, however, is nowhere near smart enough to create the automations themselves.
Such sophisticated “automate the automation” capability would also potentially resolve RPA’s brittleness issues. UiPath is working on this problem as well, implementing resiliency in its bots that can deal with infrastructure issues – the first step in implementing comprehensive bot self-healing capabilities.
It would take more sophisticated AI than we have today, however, to be able to update bots when requirements change, or when vendors update application functionality, or when data formats or schemas evolve.
We also have to look to the future to see a world when automation can drive digital transformation directly. Today, automation can transform the processes and roles within an organization, what we call digitalization. Digital transformation, in contrast, transforms the business and its strategy. (See my 2018 article where I explain the difference.)
Many of the customers at Forward IV were certainly in the midst of their digital transformation journeys – and automation was unquestionably an enabler of such transformation (or they likely wouldn’t have attended the conference).
Indeed, transforming processes and roles – digitalization – can lead to greater efficiency and morale, which in turn frees up people to work on digital transformation efforts. Today, however, don’t expect your robots to digitally transform your organization.
Jason Bloomberg is founder and president of Intellyx, which advises business leaders and technology vendors on their digital transformation strategies. He wrote this article for SiliconANGLE. None of the organizations mentioned in this article is an Intellyx customer. UiPath covered most of Bloomberg’s expenses at Forward IV, a standard industry practice.