In a hot RPA market, the 'rule of five' keeps CIOs focused on use cases

The robotic process automation market is advancing at a rapid clip. “Market growth is exceeding our greater than 50% CAGR to get from $500 million to $2.8 billion in five years,” said Craig Le Clair, an analyst at Forrester. “That’s very unusual.”

Le Clair, a speaker at Forrester’s recent New Tech and Innovation 2018 conference, said that although not a new technology, the RPA market is seeing a surge in popularity because of the struggles companies have experienced with digital transformation.

A lightweight technology that automates repetitive and routine processes, RPA works through existing desktop UIs giving CIOs who want to show some kind of digital progress a chance to work with a digitization technology that doesn’t require the modernization of any system, he said.

In response, companies like Automation Anywhere and UiPath are building in features, such as combining bots with advanced analytics, to make their tech competitive and even more relevant. Indeed, Le Clair said text analytics has been a “battleground” for the RPA market this year.

“Being able to rip through documents, rip through emails, use text analytics to categorize, summarize and basically give direction via normalized data sets to RPA bots to do things,” he said. “That’s a killer app that’s raising the value of RPA the most in organizations.”

But not every process is a good fit for RPA. Le Clair suggested companies use his “rule of five” to determine where to use RPA in the enterprise.

1. No more than five decisions. Le Clair said that RPA works well for simple applications that that operate in high volume. He defined simple to mean less than five decisions. “When you get to more than five, you’re going to need a rules engine, BPM, machine learning to help with rules,” he said. “Because those decision have to be coded explicitly in a bot — in a script.”
2. No more than five apps. RPA doesn’t rely on APIs, which means they are sensitive to changes in applications. “When applications change, they often break the bot,” Le Clair said. “This has been a huge headwind for the category.” He suggested companies keep the number of applications involved to less than five.
3. No more than 500 clicks. RPA bots record the way an employee moves through repetitive tasks – cutting and pasting or data entry keystrokes. Le Clair suggested that companies keep the number of keystrokes a bot has to master to less than 500.

“If you find simples processes like these that are in high transaction volume — that’s RPA gold,” he said. “And that’s why this category has really accelerated.”