Confluent launched Streaming Agents to enable real-time AI agents with access to up-to-date contextual data from event streams. The feature builds on Apache Kafka and Apache Flink to combine data processing with AI reasoning so agents can adapt quickly and interact with other systems. Practical applications include continuously monitoring competitor prices and automatically updating a company's product pricing. The offering includes tool calling via Model Context Protocol (MCP), secure Connections to models and vector databases, External Tables and Search to enrich streams with relational and REST data, and Replayability for development and evaluation with real data. Streaming Agents are available in open preview.
The challenge lies not in the technology itself, but in its implementation. AI agents have potential, but organizations struggle with practical application. The problem lies mainly in data complexity. "Agentic AI is on every organization's roadmap. But most companies are stuck in prototype purgatory, falling behind as others race toward measurable outcomes," says Shaun Clowes, Chief Product Officer at Confluent. AI agents are only as powerful as the tools and data they have access to.
The solution offers four core components. Tool calling via Model Context Protocol (MCP) enables agents to select the right external tool for meaningful action. Connections provide secure integrations with models, vector databases, and MCP via Flink, while sensitive data remains protected. External Tables and Search enrich streaming data with non-Kafka data sources such as relational databases and REST APIs. This improves AI accuracy and vector search applications. Replayability enables development and evaluation with real data without live updates.
Collection
[
|
...
]