Banks have been using analytics or some form of it for years to predict and manage risk. We at RBL Bank have also evolved from Risk Analytics to Risk Modelling where historical data is used to play out the likely future scenarios which then become the baseline for strategic decisions, says Topendra Bhattacharjee, Head-Digital Banking, RBL Bank, in conversation with Elets News Network (ENN).
Banking in India is reinventing itself with tech-innovations. Is India ready to embrace emerging technologies?
2019 is the first year ever where ‘fintech’ as a term was mentioned in the Union Budget. This is an indicator of just how much the financial sector, and banks being a major player in that, has moved towards adopting cutting edge technology in all facets of business, whether its customer service using bots or its paperless onboarding leveraging Aadhaar or its AI/ NLP (Natural language Processing in Artificial Intelligence) engines to deliver personalised offerings for customers.
Tech-savvy Indians are turning to the internet more and more to address their money-related queries. Also, the number of Indians having access to high-speed internet, at cheap rates, is also increasing. The Reserve Bank of India (RBI) is exploring setting up a regulatory sandbox or innovation hub for the fintechs.
More and more banks are opening up towards ‘Open Banking’ platform and collaborating with fintech partners. So it’s clear that there are a large number of favourable factors indicating a shift in that direction. Both ends of the value chain, the consumers and financial institutions, are looking for new technologies for value creation.
How significant is banking analytics in today’s era? What are the areas of concern that it is primarily catering to?
Analytics is playing a significant role in banks today across business functions. Product and portfolio management and optimisation are heavily driven by data analytics and descriptive analytical modelling. This helps in increasing the profitability of each customer while effectively managing the costs.
Advanced risk modelling helps to manage the risings costs of compliance and more importantly risks of non-compliance. Predictive analytics models using Machine Learning, Big data, Data mining helps us to effectively monitor and prevent any potential fraud or money laundering activities.
Historically, banks generated huge amounts of data which was scattered across platforms and systems. The volume and velocity of data generation have only increased over the last few years. Banks have now started deploying analytics solution to leverage this data usefully.
This data is now used for customer segmentation, at-risk prediction, customer lifecycle and lifetime value management, next best and bundled product offerings etc on the business side.
Pattern recognition and predictive models help in fraud prediction, detection and prevention. Analytics models also stress testing on the assets side to predict default and repayment risk on the book. We recently implemented a model to optimise the cash flow and refilling of ATMs which helped us reduce cash-outs in our ATMs.
What is RBL Bank’s take on banking analytics? What helps it in risk management?
Banks have been using ‘analytics’ or some form of it for years to predict and manage risk. The complexity, however, has increased manifolds in the recent years. Analytics in risk management has moved from just data gathering and reporting to predictive modelling, scenario analysis and event simulations. We also have evolved from Risk Analytics to Risk Modelling where historical data is used to play out the likely future scenarios which then become the baseline for strategic decisions.
While enterprise risks are not exactly quantifiable, analytics models help to establish and examine possible scenarios and then plan for managing those risks. Predictive models generate actionable insights for the management to plan for and effectively manage enterprisewise risks.
Advanced credit risk analytics helps us to improve underwriting decisions and increase revenues while reducing risk costs. Stress-testing of initiates helps us create likely scenarios, translating them into business parameters and quantify the risks associated and then the impact on the PnL (Profit and Loss) as well as the brand. All of this enables us to mitigate risks and swiftly capture opportunities.
Tell us about the new technology deployments at the bank.
Digital channels have become the pulse of banking customers’ interactions. With almost half of banking consumers using only digital channels for their transactions, expectations are outpacing experiences. Earlier the consumer touch-points were offline (brick and mortar) and online. Now online has diverged into various interfaces such as browser through desktop, mobile, tablet and then there are apps, chat-bots. Furthermore, a consumer being time-poor tends to access the touch points across different points of the day or on different days to complete transactions (read ‘journeys’).
Future interfaces could be WhatsApp and may be other chat-based apps which may gain prominence. In order to provide consumers with this freedom, we are building micro-services in a common layer which can now be accessed across various client interfaces and also will be able to extend the same experience when newer interfaces get added.
What are your plans for 2019 in terms of innovations?
Artificial Intelligence (AI) is fast evolving as the go-to technology. Robust and rapid processing needs, the advent of mobile technology, data availability and proliferation of open-source software offer AI a huge scope in the banking sector.
With enabling technologies becoming a lot more accessible and inexpensive, AI is now becoming mainstream, with large enterprises and start-ups looking at different opportunities.
AI adoption is still in its nascent stages, and a lot more needs to be done to realise its full potential. Application of AI and ML (machine learning) to different functions within the bank will enable personalised, contextual, efficient and predictive services to customers through Bots. Another key contribution of artificial intelligence is the recommendation engine.
Recommendation engines have been very successful and have accomplished to become a revenue driver through cross-sell and upsell in the bank’s retail portfolio. It is based on using the data from the past about customers across various offerings from a bank to make the most appropriate recommendation to the user based on their preferences.
There are many such use cases such as automation in back-office operations and processes, Fraud and Risk control, software performance monitoring. Automation of such back-end workflows will increase accuracy and efficiency in serving customers. However one of the biggest challenges is the availability of the right data.
Data is the key input for an AI, and any vulnerability arising from the source or quality of information is a serious concern in the functioning of an AI model and challenges its efficacy. Structured mechanisms for collection, validation and standardisation is crucial as a starting point.