Five ways enterprises can scale robotic process automation

A starting point for how companies can ramp up from RPA experimentation to full-scale deployment.
4 October 2019

Scaling up the RPA engine room. Source: Shutterstock

Many business enterprises are implementing robotic process automation (RPA) pilots, deploying one or more robots in their operations.

One of the biggest challenges for IT departments and shared service centers (SSCs) in pursuing RPA is creating an industrial scale. How do you go from one robot to 100 or more?

The processes, tooling, and infrastructure that organizations need in order to develop in-house robots cannot simply be incremented at scale. To be successful, enterprises must redesign their entire approach.

Here are five ways to meet the primary challenges involved in scaling RPA:

# 1 | Start looking at an operating model strategy…

…that supports the bots where they will eventually be running, beginning with the end in mind.

Once the decision has been made to scale robots — whether you manage the bots from agents’ desktop computers or a centralized production environment — the IT department should manage the configuration, software distribution and robot scripts; make support staff and technical resources available; provide and maintain security access; and track and respond to incidents. It takes time and effort to configure these processes but presenting a clear operating strategy can help the IT department move the project forward.

# 2 | Devise a business continuity plan if there isn’t one in place

If systems go down, the entire software stack and the bots need to be restarted. Some enterprises create mirrored environments so they can switch to them in the event of extended system failures.

# 3 | Leverage the cloud

The cloud is the way forward for large-scale RPA operations. For example, it makes it possible to provide extra bots with a single click to address sudden peaks in transactions. The cloud enables consumption-based, efficient models.

# 4 | Bring corporate security policies into force

Bots typically do not have an ID badge, manager, address, an office or birth date ― which may be mandatory to comply with existing corporate security policies. These policies should be updated to reflect the new complexity of requiring robots to access corporate systems that traditionally require a human being’s credentials.

# 5 | Realize that constant change is the rule rather than the exception

In some companies, IT is responsible for technical changes related to bots, and the business units (finance, human resources, etc.) are responsible for functional changes for the business process that involve bots. This approach provides more speed to resolution.

However, things become more complex when third parties, like RPA consultants, business process outsourcing providers and/or tool vendors are involved. Governance is often a key challenge in these situations. RPA projects may not progress past development and test phases due to governance roadblocks.

The preferred approach is establishing a center of excellence — typically within the enterprise shared services center organization— with responsibility for the governance, policies, and tool/vendor selection for RPA. Just like organizations design bots to interact collaboratively in cooperation with humans, providing training on how to best use bots is instrumental in ensuring the human workforce’s understanding of how the bots operate.

RPA is currently a very hot topic. While a lot of the hype is focused on enabling technologies to accelerate the development of robots, the real challenge in scaling your digital RPA workforce lies in having a flexible cloud-based platform, better operating model design and, most importantly, a better appreciation of human nature.

This article was contributed by Bruno Stefanile, Senior Manager, BPS Solutions at DXC Technology.