By Balakrishna DR
Globally, the financial services industry has proved to be an enthusiastic adopter of Artificial Intelligence (AI) driven by the availability of data and investment appetite. Creative implementation of AI by start-ups and fintechs has helped further this trend. From personalisation to customer service, fraud detection and prevention to compliance, and risk monitoring to intelligent contract documents, AI has helped banks gain better control and predictability.
Today, customers expect faster, personal, and meaningful services and interactions with their banks and little tolerance for generic unsolicited messages. Therefore, banks must leverage AI to balance the need for privacy and security with personalisation and engagement. That said, the Indian banking sector has some amount of catching up to do.
Banks must adopt new business models simultaneously to integrate AI into their strategic plans and explore the use of AI for analytics and to improve customer experience. However, reliance on legacy systems, lack of data science talent, and cost constraints have impeded seamless adoption of AI. They must focus on three key aspects:
Fraud detection: AI plays a vital role in fraud detection, given the heightened threat of cyberattacks. As per the 2019 RBI annual report, losses due to banking frauds have risen by a whopping 73.8% despite the Government’s efforts to curb them. What is more alarming is that banks took an average of 22 months between the occurrence of fraud and its detection, as per RBI data. Considering RBI’s zero-liability safety net in the event of cyber frauds, it is imperative banks adopt best-fit practices and technology levers to mitigate these risks. With adoption of real-time payments, there has also been rapid innovation in the digital fraud landscape.
Set against this backdrop, banks must deploy context-sensitive AI solutions to enable advanced and adaptive real-time monitoring of their payment networks. These AI solutions additionally leverage relevant data points to assess transaction risk, true identity-matching, and identification of complex typologies and patterns.
Digitisation of processes: The tremendous proliferation of mobile devices and the internet can be leveraged to enable the superior user experience and analytics-based functionalities that give consumers an insight into their spending patterns and provide recommendations on investment and risk profiles. For instance, digitising the KYC process to eliminate the need for physical document submission and verification is something that traditional banks still do not offer. This can be simplified by utilising AI-based computer vision technology to verify documents, Optical/Intelligent Character Recognition (OCR/ICR) technologies to digitise scanned documents, and Natural Language Processing (NLP) to make sense of them.
Decision making: AI is a great fit in areas where decisions are based on available structured and unstructured data. For example, it can help predict potential loan defaulters and offer loss mitigation strategies that will work for them. It can help determine the best time to approach a customer to sell a new product. AI-based smart environments can collate data from multiple sources and drive an inference and enable SMEs to take decisions. AI can also improve straight-through processing using Intelligent Automation to automate repetitive processes that need decision making.
Given the magnitude of the challenge, it might make sense for banks to come together to establish a consortium for knowledge sharing on AI. This would also help India’s numerous regional and cooperative banks that are behind on the technology curve. A consortium could help uplift these small banks and enable them to be integrated seamlessly into a broader nationwide secure banking network. Whichever way it happens, AI in Indian banking is only set to grow.
The author is Senior VP, Service Offering Head – Energy, Communications, Services and AI & Automation Services, Infosys