The core of any AI/ML-based service is apt data. Most of us have an understanding that banking and finance companies have more structured data compared to any other industry. To know more about India’s readiness to the adoption of technologies like AI & ML, Srajan Agarwal of Elets News Network (ENN), had a conversation with Hardik Thaker, DVP – Digital, Analytics and BI, Aadhar Housing Finance Limited.
How are financial institutions transforming in sync with a dynamically changing customer base?
India has seen massive growth in digital adoption over the last ten years due to multiple initiatives by the central government. Everyone knows what role UPI has played pre/post-pandemic in India. In Aug ’22 UPI transactions in the country were ~10.7 lakh Cr (10,72,792.68 Cr) vs 2.9 lakh Cr (2,98,307.61 Cr) in Aug ’20 which is a 259 per cent growth over the last two years.
Adoption of digital banking from T3 to T6 town has drastically increased from customers belonging to all segments – professional to non-professional and HNIs to low income. So far most of Indian financial institutions have accomplished their Digital 1.0 transformation journey. However, such an enormous population shift leads to revamping existing business strategies. In addition, fintech players (including neo banks) are contributing with their innovation to serve and understand the customer differently. We will of a partnership between large banks with these fintech startups in the upcoming days.
Financial institutions are building/ partnering with key players to build layers of alternative data sources, auto AI/ML services, process automation using advanced technology, and geospatial services on top of existing infrastructure. These layers will help them strengthen GTM, marketing, acquisition, and brand proposition.
In the next 5 years, the banking industry will be highly commoditised. For example, consider gasoline vs electric cars. Every automobile company has its own specialty which building gasoline engines, that gives them an advantage over one another. However, in electric mobility, it’s a battery cell. Building EVs becomes easy compared to gasoline cars (or any other EV vehicle). The only differentiator will be customer experience! Another best example is UPI – PhonePe, Google Pay etc. Government is constantly pushing towards standardization of banking and financial services in India. Therefore, in the upcoming years the institution which understands its customers best and provides best customer service will win!
Technologies like Blockchain, AI, Machine Learning, Internet of Things (IoT), SaaS etc. are playing a major role in multiple arenas. How do you see these emerging technologies laying the roadmap for the future growth?
I would keep blockchain aside because other than Bitcoin (or any alt coin) Blockchain
is in a nascent stage. Many private sector companies (IT/Banking) are developing solutions using Blockchain technology however the biggest challenge with it is computation.
See, the first-ever ML concept was defined in the 1950s by Arthur Samuel as “the field of study that gives computers the ability to learn without explicitly being programmed.” Over the next 6 decades, many researchers tried to build successful ML programs, but it eventually kicked off in 2000 onwards when we solved ML’s biggest challenge “Computation” – and the rest is history. I’m sure this time it won’t take 6 decades to solve the problem.
Also Read | AI & ML redefining customers’ banking experience
Now, if we discuss the global adoption of SaaS, AI/ML, and IoT, it can be attributed to innovations occurring in cloud computing. AWS pioneered this space with enormous tools readily available without worrying about infrastructure scaling and maintenance. As I mentioned previously, the companies that understand and serve their customers best will dominate the market. SaaS applications create a platform for companies to reach millions in no time. AI/ ML and IoT are considered the extended arms of these SaaS platforms. Together it builds an ecosystem to converge customer-centric business models.
How do you rate India’s readiness to the adoption of technologies like Artificial Intelligence (AI) & Machine Learning (ML) in the financial space?
The core of any AI/ML-based service is apt data. Most of us have an understanding that banking and finance companies have more structured data compared to any other industry. To some extent it is true however existing structured data does not suffice for an institution to do more. In addition, customers do not want their financial service provider
to scan their accounts regularly since that’s an intrusion to their privacy. In addition, financial institutions are also building alternate data solutions through partners or directly from government institutions – like NDSL, Perfios, GST APIs etc.
The financial industry is vast but if I speak from a credit/lending point of view – many organisations have attempted automating credit underwriting using AI/ML models on customer data that broadly includes customer financials (individual or commercial), bank
statements, credit bureau records to understand RTR. Since collaterally valuation is a tricky part, not everyone has succeeded yet. There are plenty of emerging service providers and frameworks built by the Central Government to set up seamless data exchange practices among financial service providers through AA (Account Aggregator) frameworks.
Lending companies are using advanced AI/ML techniques to optimize efficiencies and cut down costs for collections and customer acquisitions. In my option, the era has just begun – we will see more use cases in the coming future.
How will cloud-based ERP, automation, and cognitive innovation continue apace, and create opportunities for financial services?
Firstly, let us look at how the banking sector evolved. In the 1990s there were few central computers in the bank where only skilled people used to work on core banking services. Later, banks started setting up only one main frame machine at individual branches across India – all branches now connected with the bank servers. Gradually it evolved to multiple PCs and people operating through branches – now multiple employees connecting to a bank server at the same time. The 2000s net banking era began. Now customers have started accessing bank servers from fiXed locations. 2015 onwards it was the mobile banking era – now customers are accessing bank servers from any location in the world. We live in 2022 where “connected apps” is an advancement to mobile banking where third-party service providers are accessing bank servers through APIs on behalf of customers.
One must understand that these journeys required a lot of structural transformation for every financial service provider – be it CRM systems or Core Banking systems. Therefore, opting for the right ERP solutions becomes key to success for every organisation. One cannot solve linear problems with linear solutions. (increasing demand should not be solved by increasing manpower) The fact is, today’s customer wants to be treated specially and on the other side does not have patience. Here, cognitive intelligence built at the core of any product or service can reduce overhead costs of customer service to the financial service provider.
Traditional IVR systems are used to act more like robots than humans and operated on fiXed options/instructions. Modern-day chatbots, voice bots come with cognitive
intelligence that can focus on mimicking human behaviour and reasoning to solve complex problems. Best known examples are – Zomato & Swiggy’s customer service channels.
Building the right solution will improve the overall customer experience and will reduce operational costs for any financial service provider.
How inadequate infrastructural development in the housing finance department is synchronised with data analytics and AI?
As I mentioned previously, for any analytics or AI application to succeed there has to be a right data strategy. In 2022, there is no more analytics strategy – it’s just strategy in the digital world.
Now let us first understand on what basis HFCs lend money to the customer. First and very important collateral valuation. Second customer’s repayment capability. Very simple! Right. Actually, it is not. Because, collateral valuation still in India is very subjective. I’m not talking about well structured residential or commercial schemes built in Tier 1 or 2 towns.
At Aadhar our majority customer base belongs to the low income segment group living in T2 to T5 towns. Now you must relate the challenges linked with them for underwriting their loan applications. Central Government is pushing toward digital land records, deed records, etc. to bring transparency however it has just begun a few years back. API-fication data from multiple government agencies, private and public sector entities boosted digital and data analytics adoptions so far.