Mastercard has expanded its suite of artificial intelligence (AI) solutions with the introduction of a new programme aimed at developing and deploying knowledge agent tools.
The first application launched as part of this initiative is a digital assistant designed to streamline the customer onboarding process. It enables users to adopt new payment methods more quickly and efficiently.
Developed in collaboration with Databricks on its data intelligence platform, Mastercard’s in-house capability features a chat-based assistant user interface, source management, and a learning mechanism supported by human feedback.
This programme ensures that the new knowledge agents are continually trained using Mastercard’s proprietary datasets while adhering to the company’s stringent AI and data governance principles.
The initiative also eliminates the complexities associated with sourcing and vetting third-party solutions, thereby further enhancing operational efficiency.
By leveraging a large language model integrated with retrieval augmented generation (RAG) and fine-tuning, the product onboarding assistant automates routine tasks and addresses critical customer queries during the onboarding phase.
The RAG system also efficiently locates accurate information to respond to user prompts by utilising Mastercard’s existing onboarding documentation as its knowledge base.
Besides, the tool incorporates a human-in-the-loop approach, allowing feedback from subject matter experts to refine the agent’s learning process and ensure the accuracy of its responses.
Mastercard currently employs machine learning models to analyse over 143 billion transactions annually, identifying potential security and cyber threats to safeguard users and businesses.
Additionally, Mastercard utilises AI to provide valuable insights and intelligence to its ecosystem partners, improve access to credit and financial management, and promote financial inclusion for individuals with limited or no credit history.
In a separate announcement, Databricks has expanded its collaboration with business decisioning data and analytics provider Dun & Bradstreet. The partnership is aimed at providing business data and analytics on hundreds of millions of businesses and other commercial entities worldwide through the Databricks Marketplace.
