After success with RPA, ICICI Bank guns for the AI pie

As software robots in ICICI Bank process around 20 lakh transactions daily, it's moving ahead in the artificial intelligence race with experiments in advanced analytics and cognitive tools.   


Two years ago, one of India’s major private lenders ICICI Bank became the first in the country to adopt software robotics, also known as robotic process automation (RPA) on a large scale. The operations department deployed around 200 robotics software programs, which helped in processing close to 10 lakh transactions daily.

It’s been two years and the bank’s love for automation has picked up and how.

The RPA fact sheet

According to research firm IDC, RPA is a software code that automates standardized, rules-based repetitive tasks, which were traditionally done by humans.

Anita Pai, senior general manager and head of operations at ICICI Bank India, said the bank has now scaled up to 750 software robots and also doubled the number of transactions (close to 20 lakh) handled daily. “These robots are being used across different operations and LoBs including retail, wholesale banking, forex, treasury, agro and international operations. And most of the robots deployed have been developed in house,” she said.

IDC Financial Insights estimates that by 2020, RPA will be in use in 40 percent of APAC banks and insurance companies. For ICICI Bank, the benefits of using software robotics range from sharp decrease in response time (60 percent) to almost real time turnaround in some cases. “Error rates in processes have come down very close to zero, productivity has improved, and it has enabled us to become scale agnostic,” mentioned Pai.

The AI pie

Besides RPA, the bank is also leveraging AI, natural language processing (NLP) and experimenting with cognitive tools. “We are at various stages in our experiments with tech like AI, chatbots, data analytics and cognitive, which will help us further reduce turnaround time for customer processes and automate another 0.4 million transactions annually,” she said.

The bank has introduced an AI-led engine in international remittance operations, which has helped reduce the query time from 12 hours to real time. Citing an example, Pai mentioned, “Earlier, when customers would remit money from international locations to India, there was a lot of anxiety and they would write to us asking for the status.

“As it was manually handled it’d take an average of 12 hours to get back to the customer. Now, an AI-led engine reads the mail, ensures it goes through the system, gets the information and replies to the customer – all in real time.”

The bank is also using chatbots for handling customer queries – from answering questions to leading the customer to the source of information on the bank’s website. It has benefited both employees and customers.

But, what about jobs?

“We look at robotics in terms of providing the best services to our customers, minus any errors. In spite of having gone on this journey, we have actually not lost any jobs.

“Employee numbers have not dropped at all. Rather, we have made the operations more efficient and are using the workforce in areas where they can do more value added work, which means more job satisfaction,” Pai said.

Edited By : Vaishnavi J Desai