Databricks Unveils Unity AI Gateway Ecosystem for Agent Governance

At the Data + AI Summit 2026, Databricks announced a partner ecosystem for Unity AI Gateway. [1] Built on Unity Catalog, the gateway is designed to govern runtime interactions involving models, agents, and AI tools across different providers. [1] The ecosystem introduces integrations intended to manage agent identity, monitor risks, and apply runtime security. [1] Unity AI Gateway uses Service Policies to define and enforce how AI workloads operate across Databricks workspaces. [2]

Alice was announced as one of the launch partners for runtime AI security. [2] The integration connects Unity AI Gateway's policies to Alice's WonderFence platform. [2] This allows organizations to define custom controls for issues like prompt injection, PII handling, and unwanted content across multiple modalities. [2] Alice's system applies real-time AI guardrails to detect and block harmful or off-policy prompts and responses. [1]

Databricks is positioning Unity AI Gateway as a solution for enterprise AI governance fragmentation. The introduction of Service Policies and integrations with tools like Alice WonderFence suggest a focus on managing the operational risks of autonomous agents rather than just model performance. This approach directly targets the visibility and control requirements needed for production AI deployments. What the sources don't address: How the gateway integrates with non-Databricks platforms or open-source agent frameworks.