Redpanda, the real-time data platform for the agentic enterprise, is releasing the Agentic Data Plane (ADP)—a new suite for safely connecting AI Agents across the growing digital workforce.
According to the company, Redpanda ADP provides a reliable data access layer to build AI Agents with complete traceability, observability, and governance for enterprise environments.
Redpanda’s ADP is built on modern, “battle-tested” data stacks, including the Redpanda Streaming platform, the Redpanda Connect data connectivity suite, and the recently acquired Oxla standard SQL query engine and database.
“AI agents have made initial inroads to the private networks of the modern enterprise,” said Alex Gallego, CEO and founder of Redpanda. “However, organizations today don’t have an easy, integrated way to observe, control, and govern data access for agents. These goal-driven agents are reading, writing, processing enterprise data in addition to sharing it with LLMs and autonomously taking actions. The Agentic Data Plane provides the very necessary MCP aware connectivity, access control, and governance for enterprise AI. And with the acquisition of Oxla, we can now give agents a standard SQL endpoint for agents to query data in motion or at rest.”
The ADP extends Redpanda's streaming platform capabilities to encompass the full breadth of data connectivity and governance for agents, adding standards such as MCP, A2A, and SQL alongside existing standards including Iceberg, and with broad deployment models including a fully managed cloud, BYOC, or on-premise.
Key capabilities of the Agentic Data Plane include:
- Unified connectivity: Through Redpanda Connect and BYOC deployments, enterprises can bridge private and public environments, unlocking data for AI without compromising security.
- Full observability: Every agent interaction including prompts, inputs, context retrieval, and actions is logged and can be replayed for debugging, compliance, or validation.
- Federated querying and transformation: A new SQL query engine enables agents and humans to query, join, and transform data across disparate sources in real time.
- Agentic Access Control (AAC): An evolved access model designed for AI agents, providing fine-grained, narrowly scoped, short-lived, agent-bound tokens to ensure secure, auditable interactions.
The first phase of the Agentic Data Plane is available now through Redpanda Cloud and BYOC deployments. Governance and transformation capabilities will follow in early 2026, with full federated SQL querying entering beta later next year, the company said.
For more information about this news, visit www.redpanda.com.