Setting Up An RPA COE – The What, The Why, The How

Intelligent Automation and RPA are on every organization’s digital agenda now. Nobody wants to miss the automation bus and lag in the technological race. Organizations have also realized that traditional RPA solutions won’t make the cut and that they need a robust strategy while approaching this breakthrough technology. Hence, organizations are resorting to establishing RPA Centers of Excellence (CoE) to build capabilities in the space.

This panel discussion from the Zinnov Intelligent Automation Conference 2019, brings to the fore various nuances of setting up an RPA CoE. It presents a holistic view of building RPA CoEs and driving meaningful business outcomes through them. The panel comprised of Ravi Gautham, SVP, Northern Trust; Anand Krishnan, Senior Director – AI, Analytics & Automation, HARMAN Connected Services; Sourav Sahu, Automation Leader, ANZ; and Mohandoss T, Zinnov Advisor, CEO – CompassMet Consulting.

Here’s the excerpt from the discussion:

Mohandoss: First, what is the business imperative of your organization/industry from an RPA standpoint?

Ravi: Northern Trust is a large global custodian with USD 10 Trillion in assets, another Trillion that we actively manage, and business from 60 companies across the world. So, we are a huge company; but we would not be able to compete in the market from a pricing standpoint had it not been for the machines and RPA. All our competitors are in the RPA game and a lot more of them are jumping in. Efficiency and improved productivity have been key outcomes of RPA, but it boils down to competitive pricing for us. We really got into the game because we figured that we could not provide the competitive pricing for the gigantic deals, without some machines helping us along the way.

Anand: HARMAN’s primary imperative in India is to improve customer experience. Let me share an instance. One of our clients in the shipping industry approached us with a unique problem of guests struggling to locate their luggage on the cruise. When we approached the problem, we did not think of automating the process of looking for luggage. Rather we shadowed the customer and found that there were a lot of processes that could potentially be automated, including switching on the lights, TV, etc. We also saw an opportunity to reduce the amount of paper that was being used. So, we started the journey of RPA which also included AI, Cognitive Computing and other technologies, and improved the overall customer experience. We went beyond automating a finite process, that of locating luggage, and took a holistic approach to automation by providing solutions around voice-controlled technologies, NLP, etc.

Sourav: For ANZ, one of our business imperatives was cost efficiency when we started our automation journey. In a bank, there are two sides – the business or the profit-making side and the technology unit, which is usually the cost unit. If the size of the technology unit increases, it, in turn, increases the cost. So we started our journey of automation with the aim to reduce redundancy, repetitive tasks, and improve process efficiency, which further led to cost efficiencies.

Mohandoss: What is an RPA COE? How did you set it up? What was your approach?

Anand: When we started the COE to focus on customer experience, we first defined the verticals or the process areas of key focus. We focused primarily on a particular line of business, looking at sales and marketing as a function. From a skill perspective, what’s important for organizations is to have process consultants – I loosely call MBAs. What they need are people with business acumen who can clearly call out the rule book or the process framework. Further, these processes can be broken down into the pilot stage, development stage, and then the deployment stage, with each stage needing niche skills.

For instance, the pilot stage will need more process consultants who can get management buy-in. Similarly, in the development stage, you will need experts who can hone the bots, framework, AI, etc. So, from a skill perspective, I believe you need to invest more in process consultants and relatively less on tooling. You also need to ensure that you are in tune with security, compliance, governance, servers, etc. Also, don’t look at RPA as a technology solution rather a business solution, because it is the business owner who is the custodian of the bot, rather than the technology owner.

Mohandoss: So, is your RPA COE more focused on execution or both execution and consulting? Also, what is the structure of your RPA COE?

Anand: We offer both execution and consulting. When it comes to the structure of the COE, we work across four different lines of businesses. There are certain business processes where we have a centralized structure. For instance, across our IT processes, there are loads of bots that work centrally. However, our procurement processes are siloed; so we went with the localized approach in those cases.

Mohandoss: Ravi, what is your approach to an RPA COE?

Ravi: While structuring our RPA COE, we have seen the pros and cons of both centralized structure as well as the federated approach. We started with the centralized approach, where one large group addressed all the needs of all lines of businesses. But we quickly realized that it was a bit too much for one single group to handle. So, we restructured our RPA COE and built centralized teams within large businesses, instead of one central team for the whole organization. But, then again, when we are solving problems for tens of thousands of people, centralizing is a problem. To take it a step further, we thought let’s just let it mushroom wherever it can. We decided that we will encourage it, we will govern it well, and have monthly meetings so that we are all on the same page.

Currently, there are close to 8 RPA COEs that have sprung up across Northern Trust, and are innovating, making a business impact, and have quicker turnaround times. But the critical thing is that you need governance, you need an overlaying architecture and control. Also, to avoid duplication, we regularly have open and frank discussions in our monthly calls regarding new capabilities that we are working on, new tools, so that we are all up to speed, and eventually, there is no mass duplication of effort. This is how we have controlled it so far.

Another aspect is that of change management. Never underestimate the importance of change management. Some people accept change enthusiastically and quickly, while some might take time. So, the organization has to strategically go about change management and get the workforce on board the idea of change. I am a big believer in making videos of everything we do. It is more relatable, more understandable. If they see that a bot can do something in 7 minutes which a human took 4 hours to do, it will have an impact. It will also be easier to get top management buy-in.

Sourav: If an organization is starting its journey now, I would say there is no right or wrong model. It will depend on the organization’s appetite. If it is centralized but has an appetite to be flexible so that you can enable others also to do the work, organizations should go for it. But as Ravi said, governance is crucial. You need to have visibility of what is happening. It should not be a process or a body that gives approval – a yes or no – it’s more to enable the guideline processes that are in place.

Most importantly, when organizations start their RPA COE journey, they don’t think about the licensing model, the cost factors, and the amount of reuse of resources. It is important to consider these factors; however, don’t delay your journey because you don’t have a template or a catalog. Start your journey, but keep in mind the importance of catalog-based delivery or reusable components. This will help you start off in the right direction and in alignment with the business needs. I have also seen the importance of time to market in automation. You have to be up to date and use the latest tools and technologies. It will be difficult to compete with old tools if we are not smart enough or agile enough.

Mohandoss: What are some of the big trends that you see both from an internal and external perspective?

Sourav: As a bank, we are competing with some of the biggest technology giants out there. However, fortunately, there are so many advanced technologies, like the open source that comes with good security, I think it is the right era for banks to get on the RPA bandwagon. I see automation getting even bigger as an end-to-end solution. It will include process modeling, workflow, RPA, AI/ML – the entire spectrum will be automation. Customer experience will, of course, continue to be the key focus area for automation in the future.

Anand: The last five years were touch, the next five years will be all about voice and image. At least that’s what we see in our industry. Automation has to be seen as a whole rather than in silos. So, when you build your core team, don’t just look at RPA experts, but people who know process modeling, business experience, AI/ML, and other core technologies. This will improve your odds of success. Another trend that I see is that solutions are becoming vertical-agnostic. The use cases are becoming universal and in another 2-3 years, there will be templates available for solutions rather than the need to build solutions from grounds up.

Ravi: If you think through the possibilities, it could be scary, it could be endless. The machines are becoming more intelligent than ever with the increasing proliferation of ML. Clearly, we will need more transparency inside that black box, but we are at a point that with time the machines will get more and more effective. There’s one low-hanging fruit that is tough to crack is the OCR (Optical Character Recognition). If someone cracks it, it is a goldmine. There are tools but it’s not enough, and there are problems like this if cracked, they can open up a plethora of opportunities.

RPA COE – The accelerator to scale organizations’ automation journey

The industry is embracing hyper-personalization to enhance Customer Experience. Automation, AI, ML, and other digital technologies are key enablers of delivering frictionless customer experience. They can also help reduce costs and error rates while increasing efficiency. Organizations that have embarked on their automation journeys need to scale and COEs are the necessary shot in the arm to accelerate this scale.