Google Integrates Looker into Gemini Enterprise: Governed AI Analytics via Agent-to-Agent Protocol

Zero-risk pass-through security and governance

The integration is built on a zero-risk pass-through architecture :

• No data ingestion or storage — Gemini Enterprise does not ingest, replicate, or store the underlying database records when a user queries data through Looker agents. The data is processed transiently and never written to persistent storage .
• Existing permissions preserved — All row-level and column-level Looker permissions, data access controls, and security policies remain fully enforced .
• VPC Service Controls and CMEK — Gemini Enterprise supports VPC Service Controls and Customer-Managed Encryption Keys (CMEK) for added data isolation and encryption governance .
• Compliance certifications — The platform maintains FedRAMP, ISO 27001/27017/27018/27701, SOC 1/2/3, PCI DSS, and BSI C5 certifications .

The core promise: users get the power of an LLM-driven analytics assistant with no risk of data leakage, no persistent copy of sensitive records, and no bypass of existing governance rules .

New capabilities in the release

• Native interactive charts — Agents can render live, interactive charts directly within Gemini Enterprise chat, not just text responses .
• Multi-agent interoperability — Looker conversational analytics agents published via A2A can interoperate with other Gemini Enterprise agents (e.g., document agents, custom MCP-based agents), enabling cross-domain workflows .
• Insight Assistant transparency — Looker's Insight Assistant now shows users the step-by-step process it used to generate a response, including which fields and data points it referenced .