How to Build an Intelligent Process Automation Center of Excellence – Part 1
August 25, 2020 / Featured Writers, Intelligent Process Automation, Opinion Piece
This is a guest post by Vishesh Bhatia, who is part of Cognizant’s automation advisory services team in the intelligent automation practice. He has over 12 years of management consulting experience ranging across industries, with a central theme being operational excellence through automation. He led his first automation program at Cognizant 8 years ago when automation wasn’t yet a buzzword. He has worked across industries and geographies starting his career in India followed by assignments in Australia, China, the Middle East and now the United States. Today, Vishesh helps organizations at different stages of automation maturity with their automation journey – from beginners to mature teams struggling to achieve enterprise scale.
I recently participated in a webinar on a topic frequently raised by Cognizant clients with whom I consult: How do you build an automation center of excellence, or COE?
COEs are critical so companies can realize the most benefit from any transformative effort, but even more so for artificial intelligence projects, particularly intelligent process automation. AI and process automation solutions are still relatively new and not pervasive in most organizations, so a COE must nurture them, share lessons learned, and ensure discipline in how they are rolled out. Very few companies are experienced in establishing and operating an effective automation COE, which the webinar intended to address.
Afterward, it occurred to me this topic is a common pain area and would be good for a series of blog posts. My friends at Indico, which sponsored the webinar, agreed – so here we are.
Let’s start at the beginning by establishing the role a COE plays in the automation lifecycle, how it should be structured and ways it might be funded.
The anatomy of an automation COE
At its core, the following represent the key functions within an automation program:
- Intake and prioritization: The COE is responsible for managing the automation pipeline, part of which is to define a process for idea intake generated by the business, and to clearly articulate the evaluation criteria on which automation candidates are prioritized. The evaluation criteria ideally should be dictated by the objectives of the automation program which should be aligned to organizational goals like cost savings, enhanced customer experience, faster time to market, improved compliance etc.
- Automation development: This is the most critical function of the COE, in which the ideas gathered during intake turn into actual automated processes. The development team comprises solution designers, business analysts and developers. They work with business process owners to document requirements, design a future state automated process blueprint and build it using a combination of automation and AI technologies.
- Operations and support: Once an automation is built, it needs to be deployed and managed. To do that requires an operations team that will trigger or schedule automations as needed and monitor performance. This team is also responsible for any automation troubleshooting/support. Depending on the nature of the issues, the team should have a mix of business process analysts who can provide support on process-related issues, and technical analysts who can investigate any technical issues related to the tool, the code or the environment in general.
It is important to call out that there is no one-size-fits-all approach for how COEs are structured. Depending on the scale, complexity and maturity of the organization, some functions can be owned and managed by individual business teams. This leads me to my next point—the COE model.
Related Article: What is Intelligent Process Automation?
Deciding on a COE model
Depending on the maturity and scale of the enterprise automation program, organizations have three main operating model choices: centralized, decentralized or federated.
In most cases, it’s advisable to begin with a centralized model to develop expertise and, more importantly, establish standards, policies and procedures before a broader rollout of the technology. I liken this to an athlete building muscle memory with years of training before being ready for a big competition.
Automation programs have many moving parts, and all must operate in harmony for the program to function smoothly. For example, the intake form needs to align with the evaluation criteria, which needs to align with program objectives like cost reduction or customer satisfaction. Similarly, the documentation of business requirements should allow developers to seamlessly translate them into actual code without constant back and forth with the process owners. Any re-usable objects you build should be maintained in a library with a well-defined taxonomy for future reference.
These are just a few examples that illustrate why central oversight by the COE is important in the early stages of the program. Once practices are standardized and the organization gets into an operating rhythm, it can slowly decentralize the model, especially as the business begins to pursue speed and scale.
Ultimately, the goal should be to move to a federated model where most development responsibilities are distributed to individual business units while the centralized COE continues to provide guidance and maintain oversight. This acts as a force multiplier and helps organizations reach automation and ROI goals sooner than with a rigid centralized structure.
A decentralized model is usually not recommended unless the construct of the organization lends itself best to operate in a decentralized fashion with individual business units and functions making independent technical, design and funding choices, with little opportunity for collaboration and overlap.
Funding the COE
Whichever COE model you choose, the automation program, including the COE, must be funded. Organizations usually do this by using the savings derived from efficiencies gained by automating business processes. In other words, a self-funding model where the program pays for itself. Behind the scenes, most organizations operationalize self-funding programs by adopting a chargeback model where business units pay for the automations the COE builds for them.
Nidal Nasr, an end user representative on the webinar with more than 20 years of experience in the insurance industry and a wealth of automation expertise, noted it is important to consider not only the cost of funding the program, but also the cost of not funding it.
He is right, because automation use cases are no longer just about saving costs. The efficiencies introduced by automating business processes are catalysts for transformative change, leading to revenue upside and increased customer satisfaction and market share. Nearly all industries are adopting process automation solutions and, if yours is not among them, the cost of being left behind is likely higher than the cost of funding the initial program.
More to come
In coming weeks, I’ll share more takeaways from the webinar, including how to get off the blocks and how to progress to “Day 2” projects, such as those involving unstructured content. I will also cover why the COE is crucial in developing process automation methodologies and frameworks, and why change management governance is important to process automation projects. To see how Cognizant is thinking about Intelligent Process Automation. click here.
Or you can click below to view the webinar on-demand and learn about these topics now.