Confluent Introduces Streaming Agents for Scalable Real-Time Agentic AI with Confluent Cloud for Apache Flink

Confluent introduces Streaming Agents, a new capability in Confluent Cloud for Apache Flink that enables enterprises to build and scale AI agents that monitor, reason, and act on real-time data. Streaming Agents simplifies the integration of tools and data for agentic AI, accelerating the adoption of AI and enabling faster time to value and new business models. It unifies data processing and AI workflows, providing easy, secure connections to every part of a business.

Confluent, Inc. (Nasdaq: CFLT), the data streaming pioneer, has announced a significant advancement in its Confluent Cloud for Apache Flink platform with the introduction of Streaming Agents. This new capability aims to simplify the integration of tools and data for agentic AI, thereby accelerating the adoption of AI and enabling faster time to value and the creation of new business models. The introduction of Streaming Agents unifies data processing and AI workflows, providing enterprises with easy, secure connections to every part of their business.

Streaming Agents are designed to bring real-time data into data processing pipelines, allowing organizations to build agents that can reason, adapt, and take immediate action as business conditions change. According to IDC research, while many enterprises ran generative AI proofs of concept between 2023 and 2024, only a small fraction reached production, and fewer than two-thirds met expectations. This highlights the need for more efficient and effective AI integration solutions.

Key features of Streaming Agents include tool calling for context-aware automation, secure integrations, external tables and search to boost AI accuracy, and replayability for iteration and safety. By using the Model Context Protocol (MCP), agents can securely invoke external tools, such as databases and APIs, to take meaningful actions. This ensures that agents have access to the most current and complete view of data, enhancing their decision-making capabilities.

Confluent’s head of technology strategy, Andrew Sellers, noted that the ability to contextualize AI with domain-specific data in real time is crucial for enterprises to achieve meaningful ROI from AI investments. The pairing of Kafka and Flink is central to Confluent’s approach, providing durability, replay capabilities, and complex event processing for agents.

Streaming Agents are available today in open preview, allowing enterprises to start building and scaling real-time AI agents without the need for live side effects. This capability enables enterprises to iterate quickly, test agents using real data, and deploy them in production environments with confidence.

Confluent’s announcement of Streaming Agents underscores its commitment to advancing the state of real-time data processing and AI integration. As enterprises continue to invest in AI, the ability to integrate real-time data and contextualize AI decisions will be increasingly important for achieving competitive advantages.

References:
[1] https://www.stocktitan.net/news/CFLT/confluent-unlocks-scalable-real-time-agentic-ai-with-streaming-vkszn0qt7kzs.html
[2] https://siliconangle.com/2025/08/19/confluent-bring-real-time-data-ai-agents/