Admin / April 4, 2019

Revolutionizing the Payment Industry: The AI Way

Revolutionizing the Payment Industry: The AI Way

Remember those childhood days, when small bucks from relatives (that were later taken away by our parents) used to get us elated? Over time, this trend took a hit and got replaced by gifts. And then came the age of mobile wallets and all of us having smart phones- soon gift cards and online money transfer started again!

A weirdly nostalgic thing to start the blog with, but the team thought that this nostalgia would be good to set a context for how the online payment industry is here to stay, and despite being criticized over the times, the market for the payment industry will grow.

To put some facts on the table, a study by Google and the Boston Consulting Group (BCG) found that the Indian digital payments industry is estimated to reach about $(and not INR)500 billion by the end of 2020 (the next year, yes) with its GDP contribution being around 15%. The study further states that 81% of the existing digital payment channel users prefer it over cash or cheques/demand drafts even for the higher payments (more than INR 50,000).

Sounds positive right! However, we have two more things to add. More than 50% of the person-to-merchant transactions are micro transactions; i.e. their value is less than INR 100. Also, against the UPI, Bharat QR or other such channels, it is the mobile wallets which are the preferred channels for these transactions.

But this blog is not about making a case for digital payments. It’s all about telling how technologies enabling digital payments would have a lot of their concerns solved by another and yet nascent technology called Artificial Intelligence (AI).

Let us have a look at what some of these challenges for the Payment Industry are and how AI can solve them.

Fraud Detection

Remember how the last time you made an unusually big transaction and the customer service from the credit card company called up to confirm if it was you who made the transaction. You were happy that the firm cares. Imagine this happening with the kind of volumes we just spoke about in the introductory section. Checking all the transactions for their genuity manually would be either a task impossible or would require considerably high manpower; thus, skyrocketing the operational costs.

However, the usage of AI can change this data deluge challenge into an opportunity. By evaluating the data obtained over factors such as IP address, time zones, locations, transaction values, payment histories, credit limits, channels on which the transactions are usually made, etc., a risk score can be defined. Also, if there have been gradual changes in the transaction patterns; like gradual increase in the transaction amounts over time, the same would also be stored and a pattern defined.
Fraud Detection
This coupled with a CDP (Customer Data Platform) will help integrate all the information from both online as well as offline transactional histories, assigning an accurate customer transaction profile.

Thus, if there are any outliers to this profile, they would be red flagged. Further, it would not necessarily have to be addressed by a human agent. Rather, a Virtual Customer Assistant (VCA) responsible for client engagement would be able to discuss the same with the customer on their personal mobile/e-mail through an app installed or any other third-party app. With voice call facility already being integrated as an additional functionality in the VCAs, there is a possibility that even the mobile apps would not have to be installed and a simple call would do it all.

Credit Scoring:

It becomes a big pain not just for the one who has applied for a credit card, but also for the credit card firm to do all the due diligence and check if the person concerned is eligible for a card or not. As odd as it may sound considering the number of calls, we get from people selling credit cards, about 40% of the credit card requests get rejected; sometimes just because some document (for instance the Voter Id Card) could not be validated. This in turn, might persuade the firm to conclude that the applicant might default on the payments in the future.

The aforesaid is a loss for both the firm as well as the applicant. The unnecessarily tedious verification and on-boarding process results in significant number of people (to the tune of 30% over the number of people who filed for the card) dropping out.
Credit Scoring
AI based solutions, however, can analyse the data across millions of the accounts to spot payment default patterns thus giving an insight into the profile of people (or their transaction histories against possibly the salary or revenue cycle) who have high likeliness of defaulting OR the ones who have least likeliness of defaulting.

Thus, instead of generating a credit score, basis the limited factors at hand, the entire historical trends are taken into practice. How conservative we want these patterns to be, would be a supervised tech based on the values defined, but pattern identification and hence credit scoring would be an unsupervised AI output; thus, saving both time and money and speeding up the on-boarding process.

Improved Customer Service

AI can now create an overarching system by monitoring the transaction right from the time the request was made till the time the transaction is processed. By the adequate usage of ML and AI, merchants can learn the estimated success rate once the payment has been submitted depending upon the card or transaction type, transaction value, geographical location of the payee, and other such relevant parameters. By optimizing the route thus followed, the merchants can increase their authorization rate, thus reducing the number of times a transaction gets declined.
Improved Customer Service
One of the critical assistances provided to the customers can be in the form of advices to regain their financial health. For instance, Douugh, a fintech startup has developed an AI based engine to help the customers choose for loan options which would be easiest for them to repay. Similarly, by creating a psychographic profile of the customers over time, basis the interactions or transactions that have happened, AI can be used to not just provide key insights to the financial institutions but also aid the customers real time.


The beauty of technology is that it can be moulded to suit needs and if wielded well, it can create significant value for both the end user and provider. At Racetrack, we understand that the requirements of one financial service provider would be different from others and so, we offer customized solutions to suit the specific needs. Should you be one and wish to discuss it further, we are just a call away!

  • AI in Hospitality Sector

    AI in Hospitality Sector

    Learn More

  • Loan Approval An AI Powered better Customer Experience

    Loan Approval: An AI Powered better Customer Experience

    Learn More

  • Powering Sales Intelligence in Real Estate

    Powering Sales Intelligence in Real Estate

    Learn More