Google expands Gemini’s Deep Research to scan Gmail, Drive, and Chat for smarter AI reports

Google has expanded its Gemini Deep Research feature to include the ability to scan users’ Gmail inboxes, Google Drive files, and Chat conversations, marking a major enhancement in how the AI assistant compiles detailed research reports.

The update, which Google has described as one of its most requested additions, begins rolling out on desktop today and will reach mobile users in the coming days. Deep Research—an AI tool designed to create multi-page analytical reports rather than short-form answers—can now synthesise insights from a range of personal and professional data sources including emails, documents, spreadsheets, presentations, PDFs, and chat logs, in addition to standard web searches.

The integration enables users to instruct Gemini to perform complex, cross-source analysis. For example, the AI can now examine project documents stored in Drive, cross-reference related email threads, and integrate public web data or internal strategy notes to produce a comprehensive report. The upgrade effectively merges internal workspace intelligence with external research capabilities, providing users with context-rich summaries and insights.

To use the feature, users must select “Deep Research” from Gemini’s Tools menu on desktop and choose their preferred source options—Search, Gmail, Drive, and/or Chat. The AI then autonomously formulates a multi-step research plan, executes the searches, and presents a synthesised report that can be exported to Google Docs or converted into an AI-generated podcast.

Since its debut in December 2024 alongside Gemini 1.5 Pro, Deep Research has evolved through subsequent iterations such as Gemini 2.0 Flash Thinking to the current Gemini 2.5 release. The system now supports highly advanced analysis, capable of processing hundreds of pages in a single research session. It leverages Google’s million-token context window to ensure continuity across extensive tasks and operates asynchronously—allowing users to close their devices and receive a completion notification once the report is ready.

An error recovery system has also been introduced, preventing entire sessions from restarting if an individual step encounters a failure.

The updated Deep Research feature is now accessible to all Gemini users via gemini.google.com, expanding the AI’s functionality from traditional web-based queries to personalised research that draws directly from users’ digital workspaces.