Cashfree Introduces AI-Powered Agents for Automated Merchant Payment Operations

Cashfree Payments, a fintech company, offers AI agents to businesses in an effort to streamline payment processes. In an effort to diversify their income streams outside payment processing, these corporations are putting more and more emphasis on agentic use cases in India. For scenarios like retrying unsuccessful payments or following up on abandoned carts, retailers can install the AI agent named Relay.

In addition, Relay may assist with a variety of tasks, such as monitoring failed subscriptions, filing disputes, and validating cash-on-delivery orders prior to dispatch. Agents that can prod consumers who are about to back out of a purchase are helping the firm aim for a lost gross merchandise value (GMV) of INR 20,000 crore. The process can begin with a voice- or text-based engagement, or with the promise of a discount to convert them.

How Relay can Benefit SMB?

Managing money could take up to 60 hours a week for a small or medium-sized company. According to Cashfree, this can be reduced to just 45 minutes when working with their agent "Relay". According to Mayank Juneja, director of engineering at Cashfree Payments, nowadays businesses have a growth team that calls clients to find out why they left their cart or if they had problems paying. A client can be eligible for a discount from these groups. An AI agent would now be responsible for powering these functions. The agent is hosted on the infrastructure of the Bengaluru-based company.

According to the company's statement, they do not share merchant transaction data with third-party AI suppliers. The agent has other potential applications, including e-commerce, loan recovery, and others, according to Juneja. Some examples of this work include checking e-payment mandates, calling consumers to follow up, and figuring out whether it's appropriate to contact them again.

Cashfree Opening a New Revenue Stream

Financial technology companies like Cashfree have razor-thin profit margins in the highly competitive payments business. Therefore, the introduction of the agentic service creates a new source of income for them. Although the service is free to use at launch, Cashfree intends to implement outcome-based pricing down the road. The number of e-commerce carts converted utilising the agent and the number of disputes successfully addressed would serve as price factors as adoption develops, according to cofounder Reeju Datta.

Vulcan, an artificial intelligence model trained on four billion transactions, was released this month by fintech company Razorpay with the goal of increasing payment success rates and detecting fraud. The program can supposedly comprehend approximately 3,000 signals from every transaction, and it was trained using 3 trillion data points. Instead of having separate models for routing, risk, fraud, and checkout, this model combines all of these functions into a single entity, which is a significant improvement.