Banking on the Future of Artificial Intelligence for Maximi$ing Data

Ravnit Singh Kohli, Managing Director (Technology) and Head of the New York/Canada Business Unit, Synechron

Ravnit Singh Kohli, Managing Director (Technology) and Head of the New York/Canada Business Unit, Synechron

Artificial Intelligence (AI) is at the heart of a paradigm shift in the financial industry. We are witnessing a revolution as AI is having a powerful impact on banking. Most of AI-led innovation can be grouped under one of the four buckets: Customer Experience, Operational Efficiency, Revenue Generation and Risk Mitigation. Although these buckets seem different, there is some overlap.

Customer Experience: Customer centricity should be the starting point of digitally-led innovation , with an objective of building a wholistic 360-degree view of the customer. This sounds simple, yet achieving this is easier said than done. Complexities associated with fragmentation of customer data across different business groups prevail. Providing an enhanced customer experience covers multiple aspects: personalized and contextualized experiences, seamless client onboarding, unified experience across service lines, seamless self-service capabilities, Virtual Chat Bots, Robo-advisors or a recommendation engine. Also notable, recommendation models created for banking clients are much more complicated than those used in e-commerce applications. That’s because the datasets are disparate and include past banking or trading history, social events, lifecycle events, spending patterns, etc. Given the complexity dealing with individualized fragmented data, and the underlying organizational silos, optimization of the client experience is a roadmap. Banks that strategically invest into this earlier on will be the clear winners.

Operational Efficiency: Banks are always seeking to derive greater efficiencies. The advent of Robotic Process Automation (RPA), Optimal Character Recognition (OCR) and Natural Language Processing (NLP)tools has made this much easier. Shifting much of these tedious, manual tasks from human to machine, banks have significantly reduced the need for human involvement. That’s had a direct impact on everything from performance and efficiency levels to staffing and expenses. RPA solutions are generally less obtrusive within the existing infrastructure, so it’s easier for banks to see this as the path of least resistance. RPA is used to extract structured and unstructured content from emails, pdfs and documents and generate workflows based on certain inputs and specified conditions. Customer onboarding at banks, for example, is a long, drawn-out process, primarily due to manual verification across several documents. RPA can make the process much easier by capturing the data from the KYC documents using OCR. This data can then be matched against the information provided by the customer in the form. With RPA tools in the market enabling drag and drop technology to automate processes, it’s easy to implement and maintain automation workflows without any (or minimal) coding requirements.

Revenue Generation: Increasingly significant are banks’ goals of leveraging AI to generate revenue opportunities. Whether related to customer acquisition using client propensity models ,or algorithms applied to identify signals to anticipate price movements enabling more effective trading decisions, lots of banks are focused on revenue generation and increasing profitability. AI should be used as a “capability enhancer” rather than a mere automation or efficiency tool. With properly built product recommendation tools, the bank can add more revenue by cross-selling products and increasing the stickiness quotient/retention of the client, thereby reducing customer churn. Models can be built to optimize discovery of the right pricing structure for customers, while opportunities for fee increases in certain situations can be quickly identified. This aspect of using AI to “generate revenue” is not the same as, nor as easy as, the “operational efficiency bucket” where flows are well understood and documented. Gathering volumes of data and multiple datasets with the right set of attributes for revenue generation opportunities is a daunting task. That’s best left to data scientists who ensure the right set of domain features (feature engineering) is applied, while models are continuously monitored and enhanced for improved decision-making.

Risk Mitigation: Another use case for AI lies in battling financial crime. This includes identifying anomalies in datasets or spotting inconsistencies or suspicious behavior. With supervised learning, a model is trained using already-categorized data to identify potentially suspicious transactions. With unsupervised learning, the computer system identifies patterns with very limited upfront categorization and the models find ways to self-organize or cluster datasets to identify anomalies. The scope for deploying AI for risk and compliance extends beyond Anti-money Laundering (AML).Credit Risk Management is also an area where banks can predict risk-prone customers that may potentially default on loans. Other use cases include detecting fraudulent credit card transactions or mitigating security risks well in advance. This bucket also includes trader surveillance, predicting liquidity risk, and ensuring adherence to KYC process by ensuring client submitted data is compliant with regulations.

Future for Banking?

We have only scratched the surface. With the digitization of data and the increase in the volume, variety and velocity of datasets, big data offers itself to even more powerful algorithms in the future. With technology and computing power seeing an exponential growth, in the coming years banks will continue on their journey to harness AI to increase efficiency and profitability.