Cashfree introduces WebMCP-based checkout to enable AI agent-led payments
Cashfree has introduced a WebMCP-based checkout capability that allows artificial intelligence (AI) agents to interact with payment pages and assist customers in completing transactions. The development reflects the emergence of agentic commerce, where AI agents can perform tasks such as product discovery, shopping and booking services on behalf of users.
Called Checkout for Agents, the new offering is available to merchants using Cashfree’s hosted checkout, payment forms and payment links. The company said merchants would not need to modify their existing integrations to support transactions initiated through AI agents.
The capability is built on WebMCP, a web standard designed to help websites communicate available actions and functions directly to AI agents. Cashfree said the integration would allow agents to access relevant checkout information, including order details and available payment methods, without having to interpret the underlying webpage structure.
WebMCP aims to simplify AI-led checkout
Traditionally, AI agents interacting with websites have had to interpret page layouts, identify buttons and input fields, and determine the steps required to complete a transaction.
This process can involve multiple interactions with the website and increase the number of tokens consumed by AI models. Token usage can also affect the operating costs of AI agents, particularly those running on usage-based pricing models.
WebMCP addresses this challenge by allowing websites to expose available functions and instructions in a structured format. Instead of interpreting the entire webpage, an AI agent can access a defined set of actions relevant to the transaction.
According to Cashfree, its implementation can reduce token consumption during checkout by more than 40%, potentially lowering the cost and time involved in completing agent-assisted purchases.
Customers retain control over final payments
Under the new system, when an AI agent reaches a Cashfree-powered checkout page, it can access the available instructions and actions through WebMCP.
The agent can retrieve order details, review available payment methods and select a suitable option. The selected payment method can then be presented to the customer for approval before the transaction is completed.
The approach is intended to reduce errors and delays during checkout while retaining customer involvement in the final payment process.
Merchants can support agent-led transactions without integration changes
For merchants, the introduction of agent-compatible checkout infrastructure could help accommodate customers who increasingly use AI assistants to discover products and complete purchases.
Cashfree said its new capability is available across its hosted checkout, payment forms and payment links without requiring additional engineering work from merchants.
The system also provides a way to distinguish agent-assisted payment activity from transactions initiated directly by customers. According to the company, browser-level indicators associated with WebMCP actions help merchants identify when an AI agent is interacting with the checkout page.
Cashfree said the approach also provides AI agents with a defined payment route through its payment aggregator infrastructure.
Commenting on the launch, Reeju Datta, Co-founder of Cashfree, said the company had identified payment completion as a challenge in the adoption of AI-led shopping. He added that Cashfree had consulted more than 50 founders building AI agents before introducing the WebMCP-based capability.
Agentic commerce creates new requirements for payment infrastructure
The development comes as AI agents begin taking on tasks traditionally performed by users across digital platforms. These include searching for products, comparing options, placing orders and booking services.
Cashfree cited estimates that agent-led purchases could reach $1.5 trillion globally by 2030. As this model of commerce develops, payment providers and merchants may need to adapt existing checkout systems to support interactions between AI agents and payment infrastructure.
The company said Checkout for Agents is part of its broader efforts to develop payment capabilities for AI-driven applications and agentic commerce.