Confluent is releasing new features for Confluent Cloud and Confluent Intelligence. These include a managed MCP server, automatic PII detection, and support for Azure Private Link.
The updates address two specific obstacles: security risks that prevent data from entering AI pipelines and the fragmentation of tools that slows developers down. “Most AI projects fail before they reach a single customer because the data layer breaks down,” says Sean Falconer, head of AI at Confluent. “Teams have the models and the mandate, but security risks and fragmented data stop them from shipping. We’re fixing that by making the streaming layer the foundation for secure, production-ready AI.”
Central to this is the Confluent MCP server, a control plane that allows AI to control, manage, and debug streaming operations via natural language. Agent Skills add a second layer to this. They encode best practices to ensure those operations run consistently. Both features are generally available for Confluent Cloud.
Also new is automatic PII detection and redaction directly within Flink SQL, without custom development or external services. This opens the door to use in highly regulated sectors such as financial services, healthcare, and insurance. The feature is in early access for Confluent Intelligence.
For secure connectivity, Confluent now supports Azure Private Link, allowing Flink jobs to connect to Azure OpenAI, Azure SQL, and Cosmos DB via Microsoft’s private network. This keeps AI workloads off the public internet.
Integration with dbt and the IBM ecosystem
A free open-source dbt adapter brings Flink SQL on Confluent Cloud within reach of the dbt framework that data engineers use daily. Teams can define, test, and deploy streaming pipelines directly using the same dbt commands as for batch processing. Confluent is also expanding its model offerings with support for TimesFM for anomaly detection and models from Anthropic and Fireworks AI.
IBM completed the acquisition of Confluent in March 2026 for approximately $11 billion. With these new capabilities, watsonx.data delivers a real-time context layer for AI in hybrid environments via Confluent. The Real-Time Context Engine, which continuously provides up-to-date and curated context to AI applications, is now generally available.