Robotic process automation (RPA) will be commonplace in finance departments by 2020, but implementations are likely to encounter failures at three key stages, according to Gartner, Inc.

“Unlike many new technologies , RPA has the potential to deliver significant business benefits on day one,” said Johanna Robinson, managing vice president and head of finance research at Gartner.

“To deploy RPA successfully finance leaders must embrace a new mindset. Unless finance departments take a more agile approach when implementing RPA, they are likely to experience failures at each phase of implementation and won’t realize the full potential of the technology.”

RPA promises to speed up and automate routine processes while reducing the amount of errors, which in turn will enable human team members to pursue higher value tasks that cannot be easily automated.

Ms. Robinson shared the three key areas of RPA implementation where failures occur most regularly.

The Planning Stage

RPA deployments often fail to deliver on expectations because they are planned as an end-to-end process, rather than focused on a single activity within a process. A focus on mapping an entire process before automating a single activity will delay implementation significantly and, in fact, create extra work. This is because, once one activity has been successfully automated, the code can be quickly applied to other similar activities within the same or different processes.

“Finance departments can start relatively conservatively with RPA by focusing on using one bot against a number of individual activities,” says Ms. Robinson.

“It’s still conservatively possible to see an output gain of up to 10 times, compared with a full-time employee working during the same amount of time.”

In this way, organizations can reap immediate efficiency gains from RPA, without investing a lot of time planning, standardizing and implementing. Then, when gains have been realized and the pilot is working well, they can move to other similar activities and processes in an agile, iterative way.

Gartner also recommends that finance leaders focus on identifying the areas of responsibility needed to manage RPA, rather than relying on traditional, fixed roles for this purpose. Finance department leaders should account for the new competencies needed for successful RPA management, centreed around digital process design. These are largely hard-to-train competencies and organizations will likely need new hiring processes to ensure the right skills for the job.

The Building Stage

In this stage, difficulties again occur when leaders treat RPA deployment the same way as they have legacy technology projects . Traditional technology deployments have relied on a “big bang” approach, where the majority of potential use cases are mapped and tested before the project is implemented. A list of requirements is generated and vendors are asked to submit their proposals.

“You don’t need to figure out every possible use case and requirement of an RPA solution before you begin,” said Ms. Robinson. “This will just result in spending more time and money than is really needed.”

The Testing Stage

Relying too much on IT teams and vendors to identify the issues and needs for deploying robots often causes failures in the testing stage. The organization’s RPA team should take the lead in clarifying and directing support needed from IT and vendors at the appropriate times.

Gartner recommends clearly defining responsibilities for RPA activities so that the RPA and IT teams deal efficiently with issues such as setting up and monitoring robot performance, with IT providing support for the underlying technology infrastructure. Due to the highly iterative nature of RPA technology, and the unique needs of the business it addresses, the most important aspects of managing robotics requires internal steering.

“The benefits of successful RPA deployments within finance include a reduction in errors from manual work and a redeployment of full-time employees to higher value activities,” added Ms. Robinson. “But robots are only as good as the people who design and manage them. CFOs should start any RPA deployment by ensuring they understand the new agile mindset needed to implement the technology, with the right competencies in place to manage it.”