Google is turning comment threads in Docs into AI-fueled collaboration zones. As of July 28, 2026, Gemini in Google Docs can now read through your comments, summarize debates, draft replies, and even suggest edits to address a colleague’s feedback—all from the bottom bar or side panel. The rollout covers paid Workspace and Gemini subscribers first, and it’s designed to appear without you asking.

What Gemini Can Now Do in Your Docs

Google bundles the new abilities into four concrete workflows, each accessible through the Gemini side panel or the bottom bar that pops up when you’re viewing a document.

Summarize and answer questions about comments. You can ask Gemini to distill long threads into themes, flag unresolved issues, or zero in on one person’s input. Google’s own examples include “Summarize all comments from Sarah” and “What are the unresolved issues?” This is triage, not just summarization—the model reads the thread, identifies patterns, and surfaces what needs attention.

Add new comments. Gemini can post comments on your behalf. You might ask it to leave a note saying “Please verify the Q3 revenue figures, Sarah,” or tell it to review the document as a copy editor and comment on clarity, tone, and flow. The AI inserts these comments directly into the doc, but they aren’t ghostwritten—you decide whether they stay.

Draft contextually grounded replies. When you click into an open comment thread, Gemini can generate a draft response that reflects the surrounding discussion. It can also pull in information from a linked Drive file. Google says the draft is grounded in the document and existing comments, but it’s still a draft. You must review and post it manually.

Suggest document edits based on feedback. If a reviewer writes “This intro doesn’t mention the new pricing model,” you can ask Gemini to rewrite the introduction accordingly. It generates suggested edits that appear in the document, ready for your approval. This closes the loop between someone spotting a problem and someone fixing it.

All of these require edit access on the document; comment-only viewers won’t see the options. But if you have edit rights, Google may also prod you before you even ask for help.

Who Gets It and When

This isn’t a blanket update for every Google Workspace user. The new comment workflows are a Gemini subscription feature that happens to live inside Docs. According to Google’s announcement, qualifying plans include:

  • Google Workspace Business Standard and Plus
  • Enterprise Standard and Plus
  • Education Plus (plus the Google AI Pro for Education and Teaching and Learning add-ons)
  • Google AI Pro and Ultra individual subscriptions
  • Organizations with AI Expanded Access

If your plan isn’t on that list, you won’t see the feature—at least not for now. Google says the gradual rollout started July 28 for both Rapid Release and Scheduled Release domains and could take up to 15 days to complete. That means some domains may not see the new capabilities until mid-August 2026.

The Uninvited Guest: Automatic Suggestions and Privacy

Here’s the sharper edge of the update: Gemini doesn’t always wait to be summoned. Google’s docs state that Docs “may proactively suggest summarizing comments when a document is opened, or offer a generated reply when you select a comment thread.” That’s a shift from on-demand AI to an always-listening assistant that jumps in based on your cursor.

For anyone who has used comment threads as the place where unpolished, half-formed, or occasionally sensitive thoughts get typed, this raises practical concerns. Comments are often the least guarded writing in a workplace—colleagues may debate strategies, mention unverified numbers, or discuss personnel matters. Until now, those words were read only by other humans on the document, and only when they scrolled to them. With this feature, Gemini reads them the moment you click into a thread, and potentially before you even open one if it’s summarizing on its own.

Google’s Workspace Experiments help page warns users not to submit personal, confidential, or sensitive information in prompts and notes that prompts, generated text, and referenced document content may be collected for the service. The same logic now extends to comment threads. And while Google frames the feature as a productivity booster, it doesn’t notify other commenters that an AI is reading and analyzing their words. Any collaborator who wrote something meant only for human eyes should know this new reality.

The lever for users is the Workspace smart features setting. Google says the entire comment workflow capability depends on that setting being enabled. If you find the nudges intrusive, or simply don’t want Gemini processing your comment context, you can turn smart features off. But that’s an all-or-nothing toggle; it disables other AI features across Docs, Sheets, and Slides too.

For Admins: Default-On Means You Need to Decide

Google says the comment workflows are enabled by default when Gemini for Workspace in Drive is enabled. That’s a critical threshold for Workspace administrators. If your organization already lets employees use Gemini in Drive, this feature will appear in their Docs without any additional prompt. Many IT teams may not have realized they were also opting in to comment summarization and auto-reply generation.

The admin checklist is short but consequential:

  • Review your Gemini for Workspace settings in the Admin console. If that toggle is on, so is this feature.
  • Check whether smart features are permitted for end users in your domain. If they are, users can turn them off individually, but the feature still exists in the environment.
  • Identify document types where comment threads routinely contain draft strategy, legal feedback, or personnel discussions. Decide whether the time saved is worth the exposure.
  • Communicate expectations clearly: AI-drafted replies and suggested edits should receive the same human review as any other business communication.

For organizations in regulated industries or those handling sensitive internal debates, the safest path may be to disable Gemini in Drive until you’ve evaluated the implications. The feature can’t be disabled from within Docs; the off switch lives in the Admin console.

For Users: Power and Responsibility

If you’re an editor or reviewer who works with heavy comment traffic, Gemini’s new skills can genuinely accelerate your workflow. The most valuable use is triage: ask it to surface every unresolved objection or consolidate feedback from a specific stakeholder. You’ll still need to read the original comments for nuance, but you won’t have to scroll through a dozen threads to find what’s still open.

When you do use the AI to draft replies or suggest edits, remember that Gemini is generating text based on patterns, not comprehension. A generated reply can misrepresent a decision, miss the tone of an internal debate, or confidently state something incorrect. Always read it before you hit post.

There’s also a simple discipline worth adopting: if a comment thread contains anything you wouldn’t want fed into an AI model, resolve it manually before anyone else opens the document. Turn off smart features if you want to prevent Gemini from reading your comments entirely, but know that this affects your own use of AI features across Workspace.

The Bigger Picture

Google’s move mirrors a broader shift in productivity software: AI is becoming a participant in collaborative documents, not just a side tool. Microsoft has been weaving Copilot into comments in Word and Excel, and similar features are appearing in Notion and other platforms. The battle is over who can turn a document’s comment trail from a static record into an active input for AI.

For Google, this tight integration into a core workflow—reviewing and acting on feedback—makes Gemini harder to ignore. It also raises the stakes for accuracy and trust. As the feature matures, expect finer-grained controls, perhaps letting admins allow summarization but block auto-reply, or letting users opt out of proactive suggestions while keeping the side panel active.

Until then, the decision sits with admins and users: is the speed worth handing your team’s internal conversations to an AI that reads without being asked? The answer will vary by organization, but one thing is clear—the technology arrived before the conversation did.