Many transactions happen with banks and NBFCs which are completely digital. Data being the new buzz word how are institutions using and guarding this data. This was discussed by a panel of eminent guests. The discussion was led by Sumit Kumar, Director – Digital and Technology Consulting at EY India.
Piyush Gupta, Chief Data Scientist, PayU Finance
Rudra Biswal, CTO, Arthan Finance
Siddhesh Sardesai, CTO, Emkay Global Financial Services
Dipesh Karki, Co-founder & CTO, LenDen Club
Sumit: How is customer experience changing with newer technologies?
Siddhesh: It has been a truly transformative customer experience due to the pandemic. We have observed that customers have become more patient and experimental. They are ready to try newer services and products. They are also ready to spend more time with the products offered to explore the opportunities. The customer now is ready for immersive experiences with audio and video details of the products. Customer wants to be part of the development process and integrate with the organization.
Sumit: How is technology keeping up to deliver a better customer experience?
Rudra: Customer centricity has become the central theme for all organizations. In the financial sector, there is a huge change in outlook. As the new generation is more tech-savvy, they are in a way demanding services and engaging more with digital touch points. Data plays a key role and technology brings that data for customer centricity. Everyone is trying to pool data from various sources. From various touch points, data is collected which is then curated to extract information. With advancements in technology automated responses are also generated. For tier 2 & 3 cities, the front-line officers engage with customers to collect the data touch points. So that customer gets a hassle-free experience. Many fintech companies help in providing solutions for better customer experience. Our endeavour is to provide quick services with a limited turnaround time. We are trying to help realise the dreams of the customer.
Sumit: In data revolution what are the use cases being utilized across the industry?
Piyush: All financial institutions, including regulators, are using AI and ML. Process improvements are happening with the help of video KYC and customer verification has become simpler. Geo-location data helps to pinpoint the customer’s location which is simpler than reaching the customer’s house. For credit decisions, big data can be used without depending on CIBIL score alone. We are able to provide better security and prevent fraud which is transaction-specific. It also helps in the identification of irregularities at the customer level across the bank. The power of AI and ML makes these steps achievable. Machine learning is now being explored further for personalized marketing. The messages can be location specific and contextual.
For a country like India digitization is fairly a recent phenomenon. Financial inclusion is happening in a large way. Access to the internet has improved a great deal with the help of mobile phones. Penetration is still low compared to other nations. The actual number of customers who have meaningful data on the bureau is very low. There is a great opportunity to push penetration with AI and ML. This will help in realizing the dream of financial inclusion. Timely approvals of loans and other financial services are also important in achieving the financial inclusion of all.
There is a need to have constant and open dialogues between regulatory bodies and financial institutions. These new technologies have hidden impacts which need to be taken into account. For example, in one program the underlying data on which the product was developed had a bias, hence the product was found to be biased. This reiterates my point that urgent and meaningful dialogue between the regulatory bodies and financial institutions will help in reaching out to more people.
Sumit: What are the different use cases of AI or ML that you have deployed that have led to significant benefits?
Dipesh: Data can be taken as a crude form of the user’s footprint in digital space. The difference which can be created is by using the data and driving information out of it. Information is more organised architect about the user. Though the firms have access to a lot of data, the problem which is being faced is creating valuable information out of the data. This leads to a lot of use and misuse of data. This in turn is pushing the regulations. Hence, we vary in the type of data that we collect from our users directly or indirectly. We inform our customers of the type of data we will be collecting from them to build the right kind of user experience. The first use case using data helped us solve was to identify the right customer and be able to understand the right income-to-obligation ratio to make sure that we are not overleveraging the customer by providing more credit.
The next use case was to identify fraudulent customers. There were instances where reputed bank statements were attached to get credit which was detected using ML. Now that we have solved the basics of credit underwriting using ML and AL we have moved to the next stage. We are fine-tuning the customer’s borrowings. With this, we have been able to achieve less than a 4% default rate. And the default adjusted return to our end customer is between 10% to 12%. We are able to provide a variable return to our customers which gives them a good experience.
We have come up with a new product called FMPP. In this, we are leveraging data to make sure that customer is able to view real-time graph. And increase their returns or safeguard their investments. This way the customer will have a way to increase their investments. We have the scoring of loans and categories of loans and do the exposure of loans based on the performance of investors’ return. With that, we have seen around 99% return from our investors.
The next use case data that we have been able to solve is the issue of communication. We have observed that customers talk to us in vernacular language. Adding to the workforce is not the solution to the rate of increase in customers. We have introduced chatbots for more interactions with the customer. Technology is able to handle almost 50-100 customers per minute to build a good relationship with the customer. They are nudged to pay EMIs on time. We send notifications to our collection officers when the system detects some delay in payments. Appropriate action is taken based on the information received. Of course, the information is cross-checked with the interaction that happened during this time.
There has been a lot of learning from failures. Almost 60% of the data collected is useful in building ML algorithms. Some data is noise which we are trying to lower. We try not to pick all the data that is flagged. That way we want to protect data privacy.
Sumit: With additional parameters, the banks are able to extract additional data. This additional data should improvise the customer experience. But do you see any negative impacts?
Piyush: AI is a great step forward in moving towards open banking. We are talking about a single framework which has access to immense accounts and data embedded in them. The most challenging aspect will be security. Customers go back to a bank if they trust it with their money. Services offered also are built on trust. The adoption of technology to build innovative products and services keeps data security as the top priority. The cornerstone will be to develop seamless customer experiences. AI and ML are still in very nascent stages. More innovation is expected to come to build products which have trust and security inbuilt into them. This will increase the adoption of AI technology in all services.
AI also helps in better predictability as a single framework can be detrimental for a particular customer. That is assuming that data in the framework is incorrect. This brings a bad experience to that particular customer. Hence data security and quality are paramount for a single framework.
Sumit: What challenges do you foresee?
Rudra: For us, the challenge has been security around data. The issue with this system sees is that customers might be hesitant on sharing data. For example, regulatory bodies advise against sharing OTP. But that will be the basis for consent from the customer for drawing of the statement of accounts. Which is important for any lender to forward credit. We see a lot of service providers joining the ecosystem. They are not supposed to save the data and pass it on to the end user like a lender.
Sumit: How do data privacy and data management come together and deliver a better customer experience?
Siddesh: The crux of data privacy is customer consent. That also determines trust in the technology of the financial institution. There is data residing both inside and outside the organisation. It comes down to how the data inside is managed by encryption and providing anonymity. Based on this the customers share their data with the institution. How is the institution handling the outside data? And how is it being put to use?
Dipesh: Data is important and building information out of it is very important. It is equally important that we handle customer data with utmost care. As custodians of data, we have to stop leakage or misuse of data.