View: Indian banks need to get way smarter

Even as artificial intelligence (AI) is being used across sectors and disciplines, it is banking that has adopted it faster than most other sectors. Today, many banks use advanced AI chatbots to solve customer queries and provide better services. Some use robotics software in their business processes across various functions, thereby processing a significant percentage of transactions with higher accuracy. In the area of early fraud detection, AI is also extremely useful.

Visual identification and verification is another important area where banks are benefiting through AI, especially when it comes to insurance claims. Geotagging can help more banks target its customers at a granular level with a high success rate.

Indian banks are taking more operational decisions nowadays through machine learning (ML), mainly in automated loan processing, customer acquisition, credit monitoring, customer retention, cross-selling, upselling, churn analysis and delinquency prediction. Personal savings and wealth management, based on historical transaction and behavioural data of customers, is an area in which likely expenses can be forecast and, accordingly, bank can prescribe ideal savings and investments to its customers.

AI implementation will result in lower cost of acquisition, lower probability of default, stronger customer activation, higher deepening of products, lower churn rate and better compliance. All these will help banks engage with its customers from acquisition to retention and, thus, create value for the organisation.

Even though banking is ahead of other industries in analytics maturity, there is still enough room for more innovation. Early adopters of AI stand to gain a significant advantage and experience. Despite billions of dollars spent towards technology upgradation each year across the world, few banks have successfully been able to scale AI technology across the organisation. The main limiting factors are lack of a clear AI strategy and its required infrastructure.

For this purpose, an AI-first approach is needed. This will make for an 'intelligent bank', where most business decisions will be taken through mix and match of human and artificial intelligence. As a first step towards an intelligent bank based on core competency, size, business complexity, risk appetite and degree of aspiration, banks first need to devise their own strategy, the areas in which AI should be and should not be adopted.

All low-end back office-based processing jobs, for instance, can be automated, thereby saving human capital and improve efficiency with reduced turnaround time. These back office resources can be reallocated in multiple other activities. At the front office level, in areas like marketing, sourcing new business, delinquency prediction and churn prediction, a bank may judiciously use its AI capabilities wherever required based on its skill set, required tools and techniques.

As many banks in India have limited IT infrastructure and specialised skill sets for advanced analytics, banks could tie up with fintech companies, where banks will provide its domain expertise along with its wide database, while fintech will provide its latest technology. However, banks need to come up to speed, agility and flexibility with fintech. They need to invest adequately for strengthening their core technology to support various analytical tools. Data purity is another important area. The proper skill sets for converting raw data to meaningful output is a priority requirement.

Regulators of emerging technology will have to play a proactive role in framing proper regulation to balance the business interest of banks without compromising customer data privacy and security. The future lies in analytics, where early movers will have the comparative advantage to improve their competitiveness and profitability by offering intelligent banking services.