GoCardless has recently rolled out a Model Context Protocol (MCP) tool, facilitating interactions between its platform and developers or merchants through natural language thanks to an LLM.

This new tool supports two key functionalities. Firstly, it enables ecommerce partners to detail their payment needs in plain terms and receive both customized implementation advice and ready-to-use code. This reduces the necessity for extensive technical knowledge about bank payment protocols. Secondly, merchants can query their live payment data using natural language. For instance, a merchant might inquire about overdue payments on a specific day and obtain an immediate response based on their own GoCardless records.

Simplifying Integration Processes

By streamlining the integration process, MCP aims to accelerate the timeline for merchants from decision-making to live implementation of GoCardless services.

Towards Autonomous Commerce

GoCardless envisions MCP as a stepping stone towards ‘agentic commerce’, where AI-driven agents can autonomously handle transactions. The protocol is designed to ensure secure and independent interactions with GoCardless’s platform.

The company’s investments in AI extend beyond MCP, including the development of Success+, which optimizes payment retries and recovers about 70% of failed payments on average, and Protect+, a fraud detection system that thwarts risky transactions before they are processed.

MCP represents an evolution in how payment infrastructure can be accessed, moving from detailed technical documentation to more user-friendly, conversational interfaces built atop existing APIs. As the use of LLMs proliferates among both developers and non-technical business users, platforms that expose their functionalities through standard protocols could see enhanced usability and a broader reach.