Databricks is building out its Unity AI Gateway to create a more robust AI governance ecosystem. This initiative aims to address the expanding governance needs as organizations move AI from experimentation into production environments.
The expanded platform focuses on integrating a wide array of third-party solutions for AI security, identity management, observability, and agent governance. This strategy allows enterprises to leverage existing security and identity tools within their AI workflows.
Runtime Security and Guardrails
As AI agents gain access to sensitive enterprise systems, real-time protection is paramount. Unity AI Gateway now integrates with solutions from Alice, CrowdStrike, Cyera, HiddenLayer, Netskope, Noma Security, Obsidian Security, Openlayer, Palo Alto Networks, and Zscaler. These partnerships provide guardrails for prompts, model responses, and agent actions.
These integrations offer features like prompt injection detection, data exposure prevention, and content moderation. They ensure that AI interactions align with enterprise policies and security standards.
Identity and Access Governance for Agents
Governing AI agent identities and their access to resources is crucial. Databricks is integrating with Okta, Ping Identity, and Saviynt to extend enterprise identity controls to AI agents. This allows organizations to manage agent identities and delegate access consistently.