Understand this story
On November 20, 2026, Google made Semantic governance policies available in preview on the Gemini platform for enterprise agent governance, checking users’ original intent against enterprise rules written in natural language before tool calls are executed. Proposed calls are submitted through Agent Gateway for evaluation by an LLM judge; if they fail, the system replaces the response with a configurable denial message and records the ruling, rationale, and token usage in Cloud Logging. Google recommends first using dry run mode to review rulings before switching to enforcement. There are currently no other evidence-backed developments.
AI-generated from public reports
Latest development · Report
Google Launches Semantic Governance Policies for the Gemini PlatformDevelopments
1 confirmed developments
Semantic governance policies are now available in preview on the Gemini platform
Source:Google Developers Blog
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 20
Google Developers BlogFirst-partySelected
Google Launches Semantic Governance Policies for the Gemini Platform
Google has launched Semantic Governance Policies for the Gemini platform in Preview, checking both the user's original intent and enterprise rules written in natural language before agent tool calls are executed. Through Agent Gateway, the feature submits proposed calls to an LLM judge for evaluation. If a call fails the check, its response is replaced with a configurable denial message, while the decision, rationale, and token usage are recorded in Cloud Logging. Google recommends first reviewing decisions in dry run mode before switching to enforcement.
AI Trends