Google Gemini Deep Research looks set for a big update, as findings indicate it’s nearing the ability for you to custom-pick your data sources like Gmail, Drive, Search, and Chat when performing a research task. This would elevate Deep Research from a file-by-file view to actually being able to do deep cross-account retrieval, taking combined personal + work archives from a second-class corpus for AI-assisted synthesis to first-class.
What the latest Android app teardown reveals about Sources
Strings and UI elements found in a recent Google app for Android build reference an upcoming Sources feature within Deep Research. Tapping Sources provides a set of options that can be toggled on or off, such as Google Search, Gmail, Drive, and Chat, in addition to a Files button for uploading individual documents. Labs for toggling Gmail and Drive in test behavior give early hints that enabling this toggle provides Deep Research visibility across an entire inbox or Drive (versus browsing to each file).
- What the latest Android app teardown reveals about Sources
- Why inbox and Drive access inside Deep Research matters
- How this might play out for everyday workflows in practice
- How Google’s approach compares with rivals and alternatives
- Privacy safeguards and admin controls for broader access
- What to watch next as Google tests broader source controls
That contrasts with the current Deep Research experience on mobile, which focuses on picking individual files or images from Camera, Gallery, Files, or Drive. There’s also the fact that it’s possible to exclude the web entirely, restricting your analysis only to what you trust and/or own in-house if you’re concerned about misinformation or citation quality.
As with all features discovered through app teardowns, the functionality could change prior to going live or roll out in limited tests. Still, it’s an indicator Google is heading further in the direction of source-level control and indexation inside Gemini.
Why inbox and Drive access inside Deep Research matters
For knowledge workers, the best answers often hide in email chains, meeting notes, briefs, and spreadsheets — not on the public web. According to the McKinsey Global Institute, workers spend almost 20% of their time gathering and searching for what they need. Immediate access to Gmail and Drive might slash that hunt — bringing you relevant attachments, conversations, or working documents in one go.
Google’s scale raises the stakes. Gmail has more than 1.5 billion users, and Google has said that Workspace has “more than 3 billion users worldwide” (presumably including Drive) while Drive crossed the one billion user mark long ago. Even small improvements in productivity across those kinds of repositories mean big impact, particularly for teams that spend their days inside Docs, Sheets, and Slides.
How this might play out for everyday workflows in practice
Think about planning a client workshop: Deep Research could pull agenda drafts from Drive, verify dates through a Gmail thread, and cross-reference action items mentioned in Chat.
Or a researcher creating a market brief could add in an uploaded PDF along with a sweep of Drive folders and archived emails to generate just that — while intentionally leaving out the open web so as not to produce results from sources of questionable up-to-dateness.
This double-barreled approach to rights control gives you a choice. Users could tether their analysis to a core trunk of documents, then let Gemini’s AI go out and harvest additional findings strewn throughout their account.
How Google’s approach compares with rivals and alternatives
Microsoft has pushed Graph-grounded Copilot experiences that span Outlook, OneDrive, SharePoint, and Teams. Google, for its part, just got a $300 million fine in France over disputes about how it treats news publishers and the quality of search results (while complaining that France wouldn’t accept help integrating with user-level data). Bringing Gmail and Drive directly under the Gemini Deep Research umbrella would be Google’s matching answer on its home turf — more deeply integrated into the same repositories people use all day long.
Tools elsewhere, like Notion’s Q&A and enterprise search systems, try to do the same retrieval-augmented generation across private knowledge banks. OpenAI’s ChatGPT has file uploads and third-party connectors, but native access to Gmail and Drive inside Gemini could simplify permissions flows and onboarding for Workspace users.
Privacy safeguards and admin controls for broader access
More access will raise questions about stewardship of the data. Google says it does not train models on Workspace customer content without customer consent, and that Workspace provides admin policies, data loss prevention, and classification tools that could gate which sources Gemini has access to. If this is rolled out to business accounts en masse, you’d expect a focus on granular toggles, per-user permissions, and auditability.
Consumers will care about clear on-device prompts and scope-limited permissions. Users must be able to remove sensitive labels or folders, and should be able to easily turn off web results when they desire a response that is only internal.
What to watch next as Google tests broader source controls
Key signals to watch:
- Do source controls show up in Gemini on the web?
- Does NotebookLM get the same level of whole-account ingestion as the Sources feature?
- What other repositories, like Google Photos or Calendar, will join the list?
Pricing plan offerings and availability in the Workspace admin will also factor into how soon teams can start using the feature.
If the souped-up app hints at what Google will be shipping, then Deep Research could transform from a powerful file analyzer to an actual personal knowledge engine — grounded in your inbox, Drive, and day-to-day context — no scavenger hunt required.
