Ververica GmbH, the original creator of Apache Flink, today announced a new open-source sub-project of Apache Flink called Flink Agents, an embedded intelligent artificial intelligence system that feeds directly into business data flows.

Apache Flink is an open-source framework that can process streams of data with the help of a programming model that allows real-time data processing and analysis in large-scale settings. Using Flink, developers can create apps that can deal with data streams while providing high throughput, low latency and strong consistency guarantees.

Flink Agents have been designed in collaboration with companies including Alibaba Cloud, LinkedIn and Confluent Inc., along with the Apache Software Foundation, to offer a native framework that fills the gap in existing frameworks by enabling AI to operate as autonomous business intelligence rather than reactive tools.

Ververica argues that traditional AI frameworks often fail due to complexity and latency issues, such as waiting for human questions or scheduled checks. Flink Agents, in contrast, offer an AI system that continuously monitors business operations and takes intelligent actions automatically — without explicit prompts.

Flink Agents operate within Flink’s event-driven runtime, where they are built, tested and executed and are supported by Flink’s streaming engine, which provides exactly-once guarantees, fault tolerance, scalability and replayability.

The agents harness the power of Flink’s high-volume, stateful streaming of continuous event processing through processing fresh contextual data access and are designed with reliability and trust through exactly-once semantics, testing, debugging and audits.

Flink Agents can power next-generation fraud detection and real-time product personalization, giving users a significant edge. The agents can benefit multiple industries and are said by Ververica to thrive in complex environments, where multiple specialized agents work together and invoke tools to solve specific business challenges.

“By contributing to Flink Agents, we are demonstrating our commitment to making real-time AI a reality for enterprises,” said Chief Product Officer Vladimir Jandreski. “This project is a powerful validation of Ververica’s focus on advancing real-time data lakes, stream-first architectures and AI-driven capabilities that enable organizations to act instantly on their data.”

The company noted that the Flink Agents development exemplifies open-source collaboration, as through the Apache Software Foundation, Ververica, LinkedIn, Confluent and Alibaba Cloud are establishing an industry standard rather than a proprietary solution. The approach allows for enterprise adoption backed by global developer community support and validation from companies operating at massive scale.

Image: SiliconANGLE/Ideogram