Google is expanding Gemini Enterprise with new ways for teams to ground answers in company context, including Chat becoming a selectable data source inside the Gemini experience. The rollout also consolidates business-focused AI capabilities into one offering that keeps a familiar interface while adding pro-level tools and centralized administration.
Gemini Enterprise includes access to advanced models and first-party features such as Deep Research and Data Insights, while pushing further into agent workflows for employees to discover, create, share and run assistants inside a secure environment.
Key updates called out in recent release notes include:
- App-level IAM policies for fine-grained access control
- Usage audit logs available in Cloud Logging
- New data connectors for Notion and Linear in public preview
- Microsoft SharePoint connector support for analyzing XLSX and CSV content without uploading files
- Admin-controlled agent sharing, including Google-made, employee-made, ADK-registered, Dialogflow-registered, Marketplace agents
- A preview Data Science Agent focused on data wrangling, ingestion, exploration and multi-step training and inference plans
- Model availability changes, including Gemini 3.1 Pro (Preview) with an admin toggle
Editions also differ on data handling: the Starter edition can use data for product improvements with an opt-out, while Business, Standard and Plus state prompts and outputs are not used to train models for other customers.
