Google put a drafting assistant inside the blank Doc

The writer's problem

A writer using Google Docs for drafting runs into a feature that changed its own public name after launch, which matters when an older article, tutorial or writing-craft guide references a tool under a name that no longer appears on Google's current documentation. The original announcement and the help page now covering the same functionality are two different documents worth reading side by side.

What the documents show

Google's Workspace blog post dated 14 March 2023, titled about a new era for AI in Workspace, announced a first set of AI-powered writing features rolling out to trusted testers in Docs and Gmail, stating that users would be able to brainstorm, proofread, write and rewrite in Docs, and describing the workflow as typing a topic and receiving an instantly generated draft to refine. As retrieved on 16 September 2026, the current Docs help article covering this functionality is titled around writing and editing with Gemini in Docs, and it describes drafting from a prompt, quick refinements such as making text more formal or shorter, and incorporating context from Drive, Gmail and the open web, gated behind an eligible Workspace or Google AI plan. The functional description in both documents lines up closely, even though the current page presents the capability under the Gemini brand rather than the launch-era framing; that continuity is established here by comparing what each source describes, not by either source announcing a rename.

The editorial choice

For a writer or editor citing this feature, the useful lesson is that a living product's documentation is not stable across several years, even when the underlying announcement carries a clean, citable date. The operative source for what the feature currently does, which languages or plans it covers, and what data it can draw on is the live help page, not the original announcement, however well the announcement is dated. This is an editorial recommendation: check current scope against the current help page before advising anyone on what a long-running AI writing feature does today.

What stays with the author

The 2023 announcement frames the workflow explicitly as ongoing human work, stating that after a draft appears the person continues to refine and edit it; the current help page's requirement to accept or reject each suggestion individually preserves the same division of labor years later. Neither document specifies exactly which connected files or permissions a given draft might pull context from, so a writer working with sensitive or embargoed material should confirm that detail separately before drafting anything they cannot afford to have referenced elsewhere.

• Does the feature's current name in a source match what a reader would find if they searched for it today?
• Which files, chats or emails could plausibly be pulled into a given draft's context?
• Is the plan being used actually eligible for the current version of the feature described here?

The announcement and the current help page describe the same underlying drafting behavior three years apart under different names, and a writer relying on either alone would only get half of that picture.