INTERVIEW OF THE MONTH
INTERVIEW WITH MR SUDHANSHU SAWLANI,
Head of Robotics & Intelligent Automation at ING Wholesale Banking
What do you believe are the critical success factors to maximize returns from RPA for any organization?
First and foremost, picking a suitable process to be robotized. It can be a complete process or part of it. One can use a process discovery method like Eliminate -> Digitize (Automate/Robotize) -> Move -> Manage or process mining tools available in the market to identify the best candidate. Next, using the right combination of technologies (for instance API) together with RPA for optimum results is a key. It is also very important to have an efficient rollout strategy including pre-launch awareness, regular stakeholder alignment & demo sessions; last but not the least clearly updated work instructions delivered to the human colleagues working alongside the bot. A less famous but very critical element is change management, timely amendment of the bot should be done to cater to any changes in the process of the corresponding IT landscape.
What does the future of RPA hold with respect to technologies like AI, ML, NLP, Process Analytics, etc? How will these technologies align to meet the needs of business users?
I believe that RPA is to become smarter and smarter. RPA has already advanced in its possibilities, from simple desktop automation to a scheduled unattended bot runner. The RPA bots can now perform operations pretty autonomously yet within set constraints as the bot is scripted. Scripting of the RPA bot has evolved as well; from rigorous code, writing to drag and drop methods. Technologies such as NLP, Advanced OCR, Machine Learning, etc. provide a complimentary boost to the functioning of an RPA bot. Vendors in this space are opening more and more integration opportunities, mainly API driven, enabling companies to leverage on the off-the-shelf products rather than building everything in-house. To move from structured only input to unstructured data processing, RPA can use Natural Language Processing (NLP); similarly, to convert images into digital format, Optical Character Recognition (OCR) can be used. With respect to the path of RPA towards Artificial Intelligence (AI), it is evident that AI is dependent on the quality and reliable data which will be more accessible once the processes are executed by consistent performing bots. Thus, RPA as such could very well become mainstream if not so much in the forefront, however with the right combination of technologies; the future of Intelligent Automation seems very bright!
How do you go about handling challenges like change management and ensure effective change management and what are good practices around that?
Change is never easy and yet for many, it brings numerous opportunities. Implementing an RPA or any other Intelligent Automation solution is no different and can seem like a threat if not managed properly. Most humans by nature prefer a stable working environment; any new implementation needs people to come out of their comfort zone, which can be painful. Basically, next to developing the right technical solution, it is crucial to take care of an important transformational factor i.e. People. Until the benefits of the change are not made clear to people involved, there will be a lot of resistance and hesitation to adapt to the change. A suggestion would be to engage not only senior management but also the team members alongside whom the bot will conduct its operations. Many a time, these employees are the ones doing the job which would in full or partially be conducted by the bot. One way of looking at it is that the bot took over a human job, another and more rightful way is that the bot relieved human(s) from a mundane and disengaging activity, leaving them with more value-added tasks to serve their customers better.
What’s the best way to manage security risks with respect to RPA and manage bots and any kinds of cyber-attacks?
With an increasing amount of digitalization efforts including RPA implementations, it is indeed very important to pay extra attention to the security aspects and be prepared to tackle the cyber attacks. To start with, an efficient governance framework with proper roles and responsibilities is a must. Next, a strict User Access Management procedure with appropriate Segregation of Duties should be employed. For example, a bot developer should not be able to single-handedly deploy and run the bot in production; similarly, a bot operator should not be able to amend the bot directly into production. Risk assessment should be conducted on two levels: one on the RPA infrastructure level, which would apply to all bots and others for every RPA implementation. The later should be carried out in a holistic manner and not only focused on a technical solution. Passwords to the applications should be managed using Vault systems and not stored on local drives, etc. As far as possible, no employee should be able to get the passwords on behalf of the bots, unless explicit approvals are given to retrieve a password for production incident investigation. Best practices such as audit logging and code reviews should be employed. Last but not the least, automated features to determine any abnormalities in process operations should be constructed and investigated.
Can you detail some of the experiences that you have had at ING Group with respect to your own RPA journey? What kind of business benefits you have realized from various automation projects?
At ING Wholesale Banking, we have been building and benefiting from RPA for about 4 years now. In the first years, it was all about testing the waters and validating the concept of RPA at different departments and corresponding operations. For the RPA Center of Excellence, this was also about learning and establishing the processes and procedures around RPA. In the learning phase, it was observed that RPA can be built for numerous business cases but it would not return the same gains everywhere. Later, we came up with a structured intake procedure, ensuring the right problem is solved by the appropriate solution. Also, a delivery methodology to efficiently carry the idea to implementation was constructed. In recent years, more focus was given to scaling the RPA and equipping the infrastructure with other Intelligent Automation tools such as NLP, OCR, etc. for seamless integration. Regarding the benefits, the most visible return on investment has always been robo-capacity made available, measured in hours. This is simply calculated by factoring the volumes of cases the bot will process per year and average handling time. Additionally, for customer-centric processes, we have managed to gauge the customer satisfaction in the form of NPS and employee engagement scores before and after RPA implementations. We have successfully deployed numerous RPA solutions, mostly running securely unattended in our very own Private cloud rigorously and 24×7. One of the most notable development at ING WB is a smart robot, built-in combination of RPA & NLP; this bot constantly monitors mailboxes of customer support teams and when a new email is received, it reads the email, understands the context, purpose, empathy, urgency, etc., validates the customer details, creates a ticket with exact details in an incident management system and routes the ticket to the right representative to process it. This bot creates more than 4000 tickets every week per team and is easily scalable around the globe for many more teams doing such a similar job. The accuracy percentage of this bot is more than 95%. Next to this, we have several bots built for various operations in the Know Your Customer (KYC) space.