November 15, 2017      

Robotic process automation, or RPA, is still in its earliest stages, but it will be catching on swiftly in the next couple of years, say numerous experts.

Business process automation (BPA) and RPA enable companies to process millions — and in some cases, billions — of transactions, according to PNMsoft, a GenPact company. They include addressing customer requests, assigning work among systems, and handling other tasks.

Business Takeaways:

  • The global market for robotic process automation will grow to $2.46 billion by 2022, according to Markets and Markets. Research and Markets is more bullish, predicting $5 billion by 2022.
  • Many companies are evaluating automation to reduce the burden of tedious tasks on employees, but humans are still necessary for higher-value missions.
  • RPA can also be useful for workflow management and applying business knowledge.fort

Getting started with RPA

Many of the Fortune 500 have some level of RPA operating today, explained David Poole, CEO and co-founder of Symphony Ventures. The London-based managed services company today announced a partnership with software vendor UiPath.

RPA can speed up processing of tasks.

RPA could improve the experience for both workers and customers.

“Most companies have started with science projects and are still in the early investigation stages,” he said. “The opportunity is twofold. First, you reduce human labor in the process. That allows you to really work on improving the customer experience.”

By using RPA, companies can speed up processing of many tasks so that solutions can be found for internal and external customers, such as in contact-center settings, Poole told Robotics Business Review.

For the most part, today’s RPA systems handle simple, repetitive tasks, said Anthony Abbattista, a principal at Deloitte.

In a recent whitepaper, PNMsoft cited the hypothetical example of a reinsurance company that receives transactions from multiple primary insurance companies in multiple file formats and file structures. RPA enables the reinsurer to convert these different files to the company’s own system for processing.

Humans still necessary

Though automating millions of transactions for a reinsurance firm or any other company would certainly make processing quicker, RPA by itself isn’t the total solution, according to PNMsoft.

In back offices as well as in factories, automation can handle common, high-volume, manual tasks, freeing humans for more complex tasks. Some transactions will almost always need a human touch because they are exceptions that fall outside of automation’s parameters.

In the contact-center scenario, RPA can access not only a company’s internal knowledge base, but also exterior systems much quicker than a human agent can, Poole noted. Thus, a customer’s query should be handled much more quickly and thoroughly by automation.

This use of RPA would also enhance the customer experience, because it would end the offshoring of many of these tasks, according to Poole. Many of those offshored tasks today are handled by human agents who don’t speak English well, leading to customer frustration in understanding verbal communications.

RPA would be less expensive, enabling companies to invest more in improving human interactions.

Along with the human interactions at the end of such a process, PNMsoft points out that highly trained human experts are needed at the outset to properly set up the automation so that it can automate as many transactions or other processes as possible.

In addition, the initial setup can be expensive, with the cost savings coming further down the line. The more transactions RPA can handle, the quicker the payoff. Unlike humans, RPA systems can process transactions around the clock.

RPA for workflow management

Beyond simple processing, RPA provides improved payoffs for organizations that use it not only for transaction processing, but that have also added dynamic workflows with smart business rules that push the process from one step to the next level, PNMsoft said.

RPA can help with dynamic workflows.

Dynamic workflow also depends on “ad hoc automation” that offers different “decisioning” depending on variables. For example, if a credit-card applicant has credit scores above a certain level, the application is automatically approved.

If the application is below a certain score, it is automatically rejected. If it’s between the two scores, the system can call for a more granular investigation of the credit scores or forward the case for human decision making.

In this scenario, all but the most complex decisions are automated so that humans are needed for only the most complex cases. As more of these decisions are processed, the more the variables can be refined so that some of the transactions that had been referred to humans can then be completed via RPA. This can further enhances the efficiency of these systems and the ROI of the automation investment.

More About Process Automation and AI:

Speed and quality

Automation provides only improved speed of the tasks, which does little good if the tasks themselves are designed poorly, Abbattista said. “You don’t want to do automation just for the sake of automation,” he added.

By adding business knowledge and cognitive skills to RPA, “we can up the ante,” Abbattista said.

For example, advanced cognitive capability would enable RPA to handle more gray areas for credit applications, not just the automatic approvals and declines. As a result, people would have fewer exceptions to handle.

As RPA evolves, it could eliminate the need for humans to conduct an increasing number of tasks. However, experts agree that humans will still be needed for the most complex jobs.

Beyond the contact center, application processing, and similar repetitive tasks, Poole said he sees RPA moving increasingly into back office processes to replace complex legacy systems.

Process automation is also moving into financial services to help with “know your customer” and other regulatory requirements, as well as healthcare for integrating accounting, patient care, and other legacy systems.