Mastercard has expanded its suite of AI solutions with a new initiative focused on creating and deploying knowledge agent tools. According to the announcement, the first tool introduced as part of this program is a digital assistant that streamlines the customer onboarding process, making it quicker and more efficient for customers to access new payment options. Developed in partnership with Databricks, Mastercard's new AI capability is built on the Databricks Data Intelligence Platform.

The solution features a chat-based assistant interface and capabilities for source management and human-driven learning, enabling Mastercard’s knowledge agents to be continually trained on trusted internal data while adhering to Mastercard's AI and data governance standards. This approach eliminates the need for external vendor solutions, further simplifying operations. The onboarding assistant automates routine tasks and addresses essential customer questions during onboarding, using a large language model enhanced with Retrieval Augmented Generation (RAG) and fine-tuning.

By referencing Mastercard’s existing onboarding documents, RAG locates precise information needed to answer user queries accurately. Additionally, a human-in-the-loop approach incorporates expert feedback, ensuring continuous learning and improved accuracy in the assistant’s responses. Positioning Mastercard for the future The announcement continues to explain that AI, like other advanced technologies, plays a key role in Mastercard’s mission to drive economic empowerment.

Currently, Mastercard employs machine learning models to analyse over 143 billion transactions annually, helping to identify and prevent security and cyber threats. AI also aids in providing ecosystem members with critical insights, enhancing credit access and financial management, supporting financial inclusion, and detecting potentially compromised cards. Source: Link .