In my first and second blog posts of this three-part series on robotic process automation (RPA) from a practitioner’s perspective, I wrote about how an enterprise strategy can make an RPA strategy bulletproof and about the new responsibilities RPA introduces to enterprise IT teams.
Those blogs covered a lot of the ground needed to position an enterprise RPA implementation for success, but there is one more critical piece required to complete the picture: a methodology for RPA implementation at enterprise scale. That’s what I’ll talk about here.
Many RPA implementations are small, affecting just one or two processes. Ultimately, once an organization’s leaders are comfortable with it, it can grow into a strategic initiative that touches almost every part of the mission.
The combination of significant demand and the urge to reduce operational expenditures can result in inadequately designed implementations that need constant care or are short-lived. Without appropriate governance, such projects can suffer from lack of accountability and operational overhead that call into question the return on investment and the viability of the technology. Further complicating matters, enterprises may decentralize RPA project governance, leading to multiple centers of implementation, duplication of effort, and unnecessary technology investments. Even with a well-developed RPA strategy, the lack of a methodology to deliver RPA across the enterprise can result in failed implementations.
An enterprise RPA strategy demands an enterprise-ready implementation methodology. In my first post in this series, I recommended an implementation methodology focused on the Robotic Operations Center (ROC)—a unified approach to achieve strategic outcomes with automation at scale. The ROC includes experts who design and develop RPA as well as operational personnel who manage the RPA platform and automations deployed to the platform. The ROC is supported as required by enterprise subject matter experts (SMEs) from business mission and IT areas. They support RPA implementations throughout the lifecycle, not just in the design stage.
Here are five reasons enterprises should apply the ROC-based approach as their RPA implementation methodology:
1. Increase ROI through RPA expertise: Consistent implementation practices across an enterprise and centralization of investment are essential to achieve anticipated ROI. The ROC provides this necessary consistency and centralization. Enterprises typically experience high demand for RPA after initial successful pilots. The ROC is scalable to the needs of the enterprise and can be staffed to accommodate increasing enterprise demand for RPA.
2. Centralize backlog prioritization: An enterprise RPA strategy enables implementation across lines of business. With global operations, RPA may be implemented across geographies. Often, RPA scenarios are evaluated at the unit and geographic location where business SMEs are accessible, which can lead unintentionally to distributed RPA implementations and high overall costs. The ROC encourages creation of a centralized backlog of RPA scenarios collected from assessments across lines of business and geographies. This backlog is prioritized by business owners with a view toward mission excellence across the enterprise. With an enterprise view, the organization can size the ROC for economies of scale. ROI calculations become tangible, so expectations can be met or exceeded, in both productivity gains and quality of business operations.
3. Make teams accountable: The ROC team structure—teams of business and RPA experts advised as needed by mission and enterprise IT experts—is supported by a Responsible Accountable Consulted Informed (RACI) chart. The chart informs teams of their responsibilities across RPA tasks, distributes accountability for them and enhances communication among stakeholders. The ROC structure and accountability pattern enables tracking of RPA implementations and course corrections when required, leading to success and high ROI.
4. Design and implement for efficient operations: The ROC provides RPA experts who keep efficient operations in mind during design and development. Effective RPA design includes platform managed runtime configurations. Definition and utilization of configurations—such as enterprise web application URLs, email notification destinations, file system locations, etc.—are essential to successful RPA operations. RPA developers create configurable process-specific metadata including business rules that can change over time, reference data, exception handling, retry behavior, etc. The ROC makes operations personnel accountable for provisioning this metadata to RPA processes in collaboration with business owners.
5. Establish a sturdy cybersecurity posture: Appropriate credentialing of robots and assignment of enterprise roles are necessary for a secure RPA implementation. RPA platforms integrate “out of the box” with secure enterprise password vaults, which are essential for credential management. Within the ROC, RPA operations personnel are accountable for provisioning and maintaining robotic credentials in enterprise password vaults. Robots may be assigned different automations with the credentials necessary to perform each. However, provisioning too many roles to robots violates the NIST 800-123 principle of least privileges and exposes the enterprise to cybersecurity risk. The ROC engages cybersecurity experts to design separation of concerns among robots. Moreover, established enterprise cybersecurity strategies like multi-factor authentication may prove too restrictive when onboarding robots. Cybersecurity experts assist in creating mitigating strategies applicable to robots in a sturdy cybersecurity posture.
CGI helps our clients set up and operate their ROCs. Interested in learning more? Reach out to us for details. I also invite you to explore CGI’s capabilities in RPA and other aspects of intelligent automation for the federal government, and to watch our RPA video.
About this author
Sourabh Pawar is part of CGI Federal’s Emerging Technology Practice (ETP), where he leads innovation in intelligent automation. Since 2004 he has worked in solutions architecture, designing and developing large-scale, public-facing enterprise and web applications. He holds a master’s degree in Computer Science from Virginia ...