Confluent Makes It Easy to Build and Secure Real-Time AI at Scale

LONDON – Data streaming provider Confluent today announced new capabilities to Confluent Intelligence again Unified Cloud which guides how real-time, artificial intelligence (AI) applications are built and secured.

Confluent integrates the AI ​​lifecycle with developer tools already in place, integrating Apache Flink and dbt pipelines and introducing a fully managed Model Context Protocol (MCP) server and Agent Capabilities that allow AI to manage streaming tasks. Additionally, by automating personal information (PII) automation and private communication to external models through Azure Private Link, Confluent embeds enterprise-grade governance directly into data streams.

These updates remove the security and complexity barriers that prevent organizations from moving AI workloads into the real world.

“Many AI projects fail before they reach a single customer because the data layer breaks down,” said Sean Falconer, Head of AI at Confluent. “Teams have models and authority, but security risks and disparate data prevent them from deploying. We’re fixing that by making the broadcast layer the foundation for secure, production-ready AI.”

The problem is widespread according to a McKinsey a report that finds “eight out of ten companies cite data limitations as a roadblock to scaling up agent AI.” The root causes are often tied to security teams blocking data from entering AI pipelines over exposure risks, and developers losing hours switching tools to analyze and manage the data streams their AI depends on. The result is a slow, manual process that turns what should be a fast iterative cycle into a bottleneck.

Security Engine, Scalable AI

Synchronous Clouds and Cognitive Intelligence form the data-streaming foundation for production-ready AI that continuously processes historical and real-time data and delivers it as a reliable core to AI applications. New capabilities add security controls and developer tools needed by top industries.

• Natural Language Processing: Developers can use a fully managed MCP server as a control plane, allowing AI to build, manage, and configure streaming jobs using natural language. Agent capabilities add a second layer, codifying best practices and workflows so that those tasks are performed consistently and in accordance with organizational standards. Together, they enable developers to build and continuously improve real-time applications using powerful AI tools, bringing streaming into modern, agent-driven workflows. It is usually available on Confluent Cloud.
• Default Data Privacy: New built-in ML function to retrieve PII and re-protect sensitive information directly from Flink SQL, without custom code, external services, or moving data to a warehouse first. This opens up many AI use cases across highly regulated industries such as financial services, healthcare, and insurance. Available in early access to Confluent Intelligence.
• Secure Communication: Support Azure Private Link ensures that AI workloads stay off the public Internet with secure, private ways to call external models and query external tables. Now, Flink operations can securely connect to Azure managed services such as Azure OpenAI, Azure SQL, and Cosmos DB over Microsoft’s private backbone. It is usually available in Confluent Cloud.
• Unified Engineering Workflows: I a free, open source dbt adapter brings Flink SQL to Confluent Cloud into dbt, a data framework that developers use to build and manage data pipelines. Teams can quickly define, test, and deploy live streaming pipelines using the same dbt commands and project structure they rely on today. This lowers the barrier to Flink adoption and makes it easier to extend existing data workflows into real-time use cases. It is usually available in Confluent Cloud.
• Flexibility with additional model and vector database support: Confluent supports TimesFM models for robust ambiguity detection as well as Anthropic and Fireworks AI models, which developers can use directly in Flink’s processing workflows to build complex real-time AI applications. Additionally, support for vector search in Amazon DynamoDB extends the modern AI stack ecosystem.

To learn more about today’s announcements, visit the Confluent blogs to find out Confluent Intelligence again Unified Cloud. Highlights include the general availability of the Real-Time Context Engine, which continues to deliver a new, controlled context for AI applications, and new fully managed connectors to the Confluent Cloud that simplify data integration.