RPA covers HDFC Ergo’s entire value chain around the claims: Mehmood Mansoori

Automation is the ultimate game, which streamlines the business. HDFC Ergo started to identify multiple opportunities where the business processes were not so automated. Currently, the company is focusing on five to six processes where robots are deployed, such as policy issuance, claims, post-policy servicing and other parts of the internal processes.

Mehmood Mansoori, President Shared Services & Online Business, HDFC Ergo General Insurance, says, “At the policy issuance, we receive a lot of data files from partners, which contains the data which is important for us to issue the policies. But these files need to be reformatted and make them acceptable to the existing systems. We first accumulated the sufficient data, to get the right limit of integration to execute a transaction without any human interference.”

Today, robots pull-off the entire job of taking those Excel files from the 40-50 multiple sources. It brings them to the required standard format. It also checks the payment information, inaccuracies in the data, visits the banking portals, does the transaction, downloads the information and statement, does the validation, etc. And further, it places the job in the queue which can be taken forward for policy processing. In case of any issues, the robots raise a ticket with the help desk.

“We are able to address the complete chain and are receiving files with almost 99 per cent accuracy, which is far superior to any humans and the RPA solution has cut down our turnaround time. And of course, the manpower requirement has significantly reduced and has resulted in cost saving” informs Mansoori.

Similarly, on the claim side, RPA solution is processing approximately 1500 claims a day. Right from collecting the data from third parties, validating the data against the various parameters and getting them approved. Hence, HDFC Ergo has covered the entire value chain around the claims and insurance processing claim and certain processes which were outsourced earlier are happening in-house. The customers can directly visit our portal, and they can execute the transaction end to end.

The turnaround time has symbolically reduced, geared-up the customer servicing efficiencies, the quality has improved and the cost has reduced. And this has generated a notable amount of positive Net Promoter Score (NPS) from the customers and has enabled us to settle the claim frictionlessly in just a matter of hours, whereas earlier it took around 24 hours.

“Our organisation is processing the claims until 10 pm, every two hourly. Nothing is waiting for the next day in the queues, we run the scheduler every two hours, initiate the NEFT, and the money is transferred in the customer’s bank account” he says.

Training the robots
Can the robots learn from their own? Currently, they are operating on the standard rules, all their decisions are data-based. For example, if there is a customer email. Can the robot interpret those emails and draw the meaning? One is the identification, categorisation of the customer’s demands. But, we are looking to send a resolution, not just the response?

Mansoori adds, “Unfortunately, as an insurance company, whatever I say in writing or verbal is a legal word and one has to abide by it. Hence, can a robot read that email, contextualise it and give an appropriate solution to the customer? After we went live with the solution, close to three per cent of our emails are responded by the Watson technology and gradually, we are seeking to jack-up numbers close to 50 per cent. We have learned the technology and we have developed our own homegrown robots, based on the open source technology.”