RPA: flash in the pan or here to stay?

RPA hype has hit an all-time high, but it needs to evolve if it is to be truly useful, argues Sanjay Srivastava

The hype around robotic process automation (RPA) freeing up staff for more sophisticated roles by taking over the repetitive, time-consuming tasks has been widespread for a while and the valuations of RPA companies have also hit an all-time high. But will the technology and providers continue this momentum? Is RPA just a flash in the pan or here to stay? What does the future hold? I recently discussed these questions and other various points on RPA with four colleagues and partners:

We came out of the conversation with some good insights into the reality and future of RPA. Here are three takeaways:

1. There are great opportunities ahead, but RPA will need to evolve

Currently, RPA effectively automates repetitive, highly structured, and high-volume processes and is where we see heavy investments. As a result, solution providers have seen significant growth in recent years. As Arvind from KeyBanc Capital Markets noted, "All are growing well over 100 percent. It seems like there is no end in sight. In turn, we project the industry to grow exponentially in the next two to three years as well."

What will happen when we have automated all the repetitive processes and run out of tasks for the bots? RPA will need to evolve from its current state, where bots simply copy what people do on a user-interface level, into more cognitive services to meet its growth potential.

Craig from Forrester was quick to note "RPA has to climb the cognitive scale to be a long-term, viable market. We're going to have to start looking at making the bots more intelligent so that we move out of a script to a more dynamic environment, which will probably mean adding machine learning. When we talk about a $2.9 bn RPA market in 2021, we assume there will be this progression in intelligence in RPA."

2. RPA is not a destination, it is part of greater journey

Many implementations focus heavily on just getting bots up and running, so much so that deployment becomes the ultimate goal. Some think they will then see immediate results in the way of full-time equivalent savings. But there is a lot more involved to get the bots set up properly, including considering what processes are ideal for automation and whether some processes need to be re-engineered prior to automation. As my colleague Riju shared, "Some see RPA as a wonder drug that works overnight. But that's not true. RPA is more like vitamins. You have to think about how to take your vitamins on a daily basis. It takes a lot of work."

Riju continued, "This difference in approach is what separates those who are forever in pilot mode and those who are industrializing RPA. The successful ones think big, but start small. They look at their end-to-end processes and identify opportunities to automate work across multiple process steps and multiple people. Then, they figure out how to get it going."

In the digital journey, we must think pragmatically, including selecting the right processes, managing the change among current employees, and having the proper controls in place to oversee bots in the new digital workforce. Those of us who approach RPA as part of this journey are more likely to get ahead than those that think of it as a one stop.

Mondelez is doing just this; during our conversation, we were fortunate to have Puneet share thoughts from his own journey implementing RPA within Mondelez. "For us, the problem we want to solve remains the same: the standardization and simplification of processes. RPA is delivering that value. That automation is real. The development cycle is short. It is very determined: drive a process at scale, simplify that process, and eliminate waste from that process. That is what we work on day in and day out. We are experimenting and learning to change our mindsets as operators, letting the bots run and work. And now, we are beginning to look at dashboards, which can tell us things like uptime, digital hours, and business exceptions. We want to feed that information back to our business."

3. RPA opens the door to a more intelligent, automated future

RPA is a great solution for its intents and purposes—basic task automation. But it won't be able to address more complicated, cognitive processes. For instance, it can't deal with unstructured data without help from natural language processing. And it will undoubtedly run into a multitude of business exceptions where people will need to step in.

For RPA to progress and be as relevant in three years from now, we must think about making these processes work in an automated fashion with all of these anomalies. This will mean incorporating other digital technologies to enhance its cognitive abilities; RPA is a gateway to intelligent automation, or as Riju put it, "a bigger, cognitive automation window."

Craig from Forrester sees a similar, promising future for RPA's intelligent progression. "We'll have areas like text analytics looking at unstructured content. We'll have chatbot interaction with the bots. We'll be able to have employees text a bot to have actions occur. We will see IoT applications, with sensory data coming from machinery getting massaged with predictive analytics, then firing actions to the bots." As we move towards this intelligent, automated future, we will see such technologies overlap and coalesce.

So, is RPA just a flash in the pan or here to stay? Well, it is here now, and it has enormous potential for growth, but only if it can evolve beyond basic tasks into more cognitive processes. For implementations, there is considerable work to be done with establishing its stability, proper use among current processes, and managing the change across people and operations. Ultimately, if we can get RPA right, it can be a good way to start making steps towards a broader, digital journey.

Sanjay Srivastava is chief digital officer at Genpact