Visa Intelligent Commerce on AWS: Agentic Commerce Revolution

Introducing Visa Intelligent Commerce on AWS: Enabling agentic commerce with Amazon Bedrock AgentCore

Visa and Amazon Web Services (AWS) are pioneering a new era of agentic commerce by integrating Visa Intelligent Commerce with Amazon Bedrock AgentCore. This collaboration enables intelligent agents to autonomously manage complex workflows, such as travel booking and shopping, by securely handling transactions and maintaining context over extended interactions. By leveraging Amazon Bedrock AgentCore’s secure, scalable infrastructure, these agents can seamlessly coordinate discovery, decision-making, and payment processes, transforming traditional digital experiences into efficient, outcome-driven workflows. This matters because it sets the stage for more seamless, secure, and intelligent commerce, reducing manual intervention and enhancing user experience.

Agentic AI is revolutionizing industries by enabling autonomous, outcome-driven workflows. Unlike traditional AI systems that are limited to answering questions or providing suggestions, agentic AI introduces intelligent agents capable of reasoning, acting, and collaborating to complete multistep tasks on behalf of users. This transformation is evident in sectors like travel, healthcare, and banking, where agents can autonomously manage research, planning, and execution with minimal human intervention. The payments industry is also experiencing a significant transformation as agentic commerce changes how transactions are initiated, authorized, and completed. This shift is reminiscent of the early 2000s ecommerce revolution, which reshaped customer expectations by making digital checkouts more seamless.

The collaboration between Amazon Web Services (AWS) and Visa is a significant step toward enabling agentic commerce. Visa Intelligent Commerce, launched in April 2025, allows developers to connect their agentic payment applications directly to Visa’s payment network using natural language commands. This initiative provides the necessary support for building network-agnostic agentic commerce flows and ensures secure communication between agents and merchants. By using Amazon Bedrock AgentCore, Visa aims to provide a secure, scalable foundation for developing intelligent commerce solutions. This partnership makes it easier for enterprises to incorporate payments into agentic workflows, ultimately transforming the consumer and business-to-business (B2B) payment experience.

Amazon Bedrock AgentCore is the backbone of these agentic commerce experiences. It provides a secure, serverless hosting environment for AI agents, ensuring sensitive data remains protected throughout workflows. The platform supports long session durations and large context payloads, enabling agents to maintain context across complex tasks like multiday trip planning or comparison shopping. Additionally, Amazon Bedrock AgentCore Identity offers secure authentication for user sign-in and access to various endpoints, while the Gateway provides governed access to tools and MCP servers. This infrastructure allows agents to handle sensitive interactions at scale, ensuring compliance with regulatory requirements and maintaining transparency through observability features.

The reusable supervisor architecture offered by Amazon Bedrock AgentCore is a key advantage for developers. This modular approach reduces development overhead by allowing the same supervisor infrastructure to be deployed across different agentic commerce scenarios, such as travel booking and shopping. By reusing the supervisor agent, developers can maintain consistent user experience patterns while adding new specialized agents for specific domains. This flexibility and consistency are crucial for creating seamless, intelligent, and secure commerce experiences. As agentic commerce continues to evolve, the collaboration between AWS and Visa sets the stage for a future where digital interactions are more connected and intuitive, ultimately benefiting both consumers and businesses.

Read the original article here