Governance of an RPA project – Beyond traditional project management

The rapid and widespread adoption of software bots under Robotic Process Automation (RPA) has accelerated the digital transformation agenda of industries globally. On one hand, it has freed the workforce from mundane swivel-chair operations, and, on the other, it has allowed people to focus on critical activities of high value.

With continuous advancements over the past few years, RPA platforms today are incorporating /embedding Artificial Intelligence techniques including Natural Language Processing (NLP), Computer Vision (includes ICR) and speech recognition.

As a result, it has fostered automation of a wide range of processes across functions, such as cash application, employee onboarding, interview updates, tax filing, tax reconciliation and invoice reconciliation to name a few.

Realizing enterprise value through RPA

With businesses achieving operational efficiency and significant productivity gains along with speed, quality, scalability, compliance and performance improvements, the role of RPA in the creation of business value is incontrovertible.

According to Forrester Consulting, 86% of organizations that have adopted RPA acknowledge an increase in efficiency; 67% report having achieved deeper insights into customers; 57% confirm having improved their customer service; and an equal percentage attest to having boosted their employee engagement rates.

Governance: Pivotal to RPA implementation

As enterprises advance their RPA footprint, there arises a need for robust cross-platform governance across systems and people, with a concrete operating model and clearly structured guidelines. This is vital for several reasons, such as:

▪ Accuracy: Although bots are programmed to run in a particular way, they need to be monitored to ensure the accuracy of the tasks performed by them. A governance model will include a checklist before a robot moves into production.

▪ Security: Since bots may log on to several applications as part of their tasks, access controls need to be created, granted and monitored as per the need.

▪ Legacy systems: Enterprises are often saddled with legacy systems that don’t integrate with the RPA applications. A governance mechanism can help overcome this issue.

It is governance that sets RPA implementation firmly on its rails. An effective governance model assigns roles, accountability, and access; ensures collaboration and communication between units; provides performance and productivity metrics and prescribes guidelines as well as templates for assessment, design, development and deployment of robots, task prioritization, change management issues and associated risks.

A governance model in RPA also establishes clear robotic process automation standards, procedures, and policies, along with governing bodies and escalation paths. It helps conform to compliance regulations, information security requirements and regulatory standards.

RPA leaders need to expand the scope of their governance model to fully reap the benefits of RPA and mitigate the risk of its failure. The model should ideally call for agreement on rollback and testing procedures for application changes, institute tech governance for the integration of bots and provide access authorization.

This would instil operative governance that would take into account real-time updates for bots and policies related to day-to-day operations. RPA operating models need to be light and federated. They should generate an audit trail, highlighting a change or decision taken by a bot.

The model would be further bolstered by regulatory governance, ensuring that the enterprise adheres to the external regulatory environment as well as internal Risk & Control Matrices (RACM) and IT general controls. The model should also include a cyber governance mechanism that protects bots and critical data from cyber threats.

Needless to say, none of this can happen without an effective integration of the human element in the form of supervision of systems. Organizations need to put in place a central governing body, commonly known as a Centre of Excellence (CoE), to validate and approve any system change in case of processes that have been automated.

Centre of Excellence for RPA governance

A governance model without a central authority consisting of humans may throw the RPA system into jeopardy. Moreover, organizations may deploy multiple automation, yet follow a siloed approach, with limited involvement of IT or central authority, thereby compromising on enterprise security and scalability.

Also, in the process of automating simple processes without involving the IT team, much of the complex processes may be excluded from automation, thus adversely affecting productivity. Different teams within an organization may engage with different RPA vendors, and this could lead to undue vendor dependence and scalability issues.

If diverse RPA systems are put in place without carrying out corresponding changes in the underlying systems, it would negate the benefits of implementing RPA.

A CoE overcomes these challenges by ensuring that the RPA program is aligned to business priorities, and has buy-in from stakeholders. It also keeps the roadmap on track and monitors as well as suggests course correction as required.

This is best achieved by following certain industry best practices:

Initiating and planning

This stage begins with articulating the tangible and intangible objectives as well as the project outcomes. Deciding the prime objective would help allocate resources efficiently—for example, if a quick operational TAT is a prime objective, then one may avoid a long series of testing cycles.

On the contrary, if high accuracy or high scenario coverage is a quality target, then one would need to put resources for extensive testing cycles.

Executing

Creating and using a standard RPA coding framework across the organization allows automation teams to create reusability and economies of scope. A standard coding framework not only helps the team with the reusability of solutions but also eases the integration of RPA solutions with organization-wide systems and applications. The risk of not doing this is that the project could falter under cost burden and technical debt.

Monitoring and controlling

A deployment strategy for various software bots needs to be formalized during the early stages of a project. Otherwise, the project may see numerous monitoring and control issues in the post-production stage.

Unlike in a typical software development cycle, RPA demands the development team to continuously monitor the performance of the bots. The deployment strategy should necessarily answer certain standard questions, such as where to deploy—on the server or on the user system? How to deploy—standalone application executables or integrated nodes?

Answering all these questions requires inputs from multiple stakeholders like internal IT teams and business owners, which may sometimes make the process of decision-making rather time-consuming. Nonetheless, it is essential to ponder on these considerations.

Conclusion

Formal business handover remains an important aspect of any project management process. Different from traditional software projects, where a final handshake between the development team and business users along with documentation on how to use the software functionalities, is a key last step, RPA projects typically do not require that since most of the bots perform activities without any manual intervention ( via unattended bots), running in the background most of the time.

However, it is important for the RPA business user team to know how to capture the outcomes of the automated process so as to ensure they can feed/utilize those outcomes in further systems if needed.

It is important to note while a user guide may not be required for an unattended bot but documentation of process steps, business logic for routing, exception handling is important for future maintenance. Responsibilities have to be defined to review daily bot production, throughput, SLA performance, and any deviations that have to be analyzed and fixed.

To summarize, complex and repetitive processes, if automated within the framework of a sound governance model, become empowered with the ingredients of success. Although, the ways and means of managing an RPA project are not drastically different from a traditional software project management, but the differences, as discussed above are significant to ensure minimized complexities, better stakeholder management and smooth governance of RPA projects.

Manish Sharma, Senior Project Manager, Optum Global Solutions