Gmail Live, Docs Live and Keep Live: Google adds vocal Gemini to its Workspace applications
Google is rolling out a new conversational layer in Gmail, Docs, and Keep with three features named Gmail Live, Docs Live, and Keep Live. Introduced in May during Google I/O 2026, these tools now allow users to speak directly to Gemini to find information in their inbox, compose documents, or take notes without using a keyboard.
This evolution goes beyond simple voice dictation. Google aims to transform its productivity applications into interfaces capable of understanding intent, leveraging context stored across multiple services, and performing tasks based on natural conversations.
Gmail Live now allows users to “speak” to their inbox
In Gmail, traditional keyword searches are replaced by a more natural interaction. With Gmail Live, users can directly ask Gemini questions about their inbox contents and continue the conversation as if they were speaking.
Rather than needing to remember the exact name of a sender or a phrase in a message, users can request information as they would verbally. A real-time transcription of the conversation appears during the exchange, allowing users to follow and verify their requests.
This approach is particularly beneficial for users with years’ worth of emails. Gmail has had a powerful search engine for some time, but constructing the right query can sometimes be necessary. With Gemini, Google aims to shift this complexity to the model: the user simply describes what they are looking for, and the AI interprets the request.
The true advantage will come from the ability to understand complex contextual questions, not simply converting a natural phrase into a classic search.
Docs Live starts from a conversation to produce an initial draft
Docs Live applies the same principle to document creation. Users can verbally explain what they want to write, allowing Gemini to produce a first version based on their input.
This feature is not limited to the information contained in the open document. Google indicates that Gemini can draw on elements from Gmail, Drive, conversations, and the web to enrich the draft when relevant.
This connection fundamentally changes the role of dictation. It is no longer just about transcribing verbatim what the user says, but about transforming an intention into structured content by automatically seeking out the necessary context.
A user can thus describe the document they need, specify their objectives, and let Gemini retrieve the information already available in their Google environment before constructing an initial version.
For Workspace, this is a logical evolution: the value of Gemini lies less in its ability to write a generic paragraph than in its access to professional context already dispersed across multiple applications.
Keep Live becomes a fully vocal note-taking app
With Keep Live, Google adopts a more spontaneous approach.
The app can serve as a voice notepad where users can jot down ideas without needing to organize them immediately. Gemini can then structure this information or generate certain content from an oral request.
Google particularly highlights the example of a list of ingredients for making shakshuka. The user can ask for what they need and then obtain a directly usable list in Keep.
This simplicity aligns well with the traditional use of the app. Keep has always been designed for quick notes, lists, and information that one wants to capture before forgetting. Voice further reduces friction by eliminating the typing step.
The potential becomes even more interesting when Gemini starts linking this note to other information or automatically restructuring a raw idea into usable content.
Google transitions from dictation to conversational action
These three features illustrate an important distinction between voice dictation and conversational interfaces. In traditional dictation, the system converts speech to text. With Gmail Live, Docs Live, and Keep Live, this transcription is merely the first step.
Gemini must then understand the intent, identify the necessary information, and perform an action. A voice request can thus trigger a search in Gmail, consult a Drive file, or create a structured document. It is precisely this layer of orchestration that is currently appealing to major players in AI.
Users no longer need to understand the technical organization of the applications. They express their goals and let the assistant determine where to find the data and how to use it.
Voice gradually becomes a common interface at Google
This launch is not unexpected. Google has been ramping up voice functions in its applications and systems for several months. In April, the Gemini app for Mac introduced dictation usable across multiple applications. More recently, the Pixel 11 featured Rambler, a dictation function powered by Gemini that can automatically clean up speech by removing certain hesitations and filler words.
Google also introduced Gemini 3.5 Transcribe in August, a model specifically designed for speech-to-text tasks.
Gmail Live, Docs Live, and Keep Live are thus the application part of a much broader strategy. Google is simultaneously developing voice recognition models, interfaces, and the necessary integrations to make speech usable throughout its ecosystem.
Context access becomes Gemini’s true advantage
The strategic point may not be voice recognition itself. Modern models already know how to transcribe speech with high accuracy. What differentiates assistants more is their ability to access relevant context.
Google holds a considerable advantage here through Workspace. For many users, Gmail contains professional conversations, Drive holds documents, Calendar contains agendas, and Docs includes ongoing work. An assistant authorized to navigate through these services can provide much more useful responses than an isolated chatbot.
Docs Live perfectly illustrates this logic: users no longer need to provide all information in their prompt because Gemini can seek it from the services it has access to. The trade-off is clear: the more the assistant can consult data, the more important permission management and transparency about the sources used become.
An interface particularly suited for mobile
Google is initially offering these functions on iOS and Android, exclusively in English. The choice of mobile is coherent. It is precisely on smartphones that voice can deliver the most significant comfort gains, especially when drafting lengthy texts or constructing complex searches is challenging with a touchscreen keyboard.
Talking to Gmail to find a message or dictating the structure of a document can thus become much more natural while on the go.
However, the experience will need to manage the context of use correctly. A voice function is practical in a closed office or at home, but much less so in transport or public spaces where privacy becomes an issue. Voice will not necessarily replace the keyboard, but it could become a much more credible complementary interface.
Google AI subscriptions determine accessible features
The rollout is currently limited to Google’s paid offerings. Gmail Live is available to Google AI Plus, Pro, and Ultra subscribers. Docs Live and Keep Live require a Google AI Pro or Ultra subscription. Google also plans to extend these functions to professional Workspace clients soon.
This segmentation indicates that advanced conversational features are becoming a standalone monetization argument. The traditional search engine, input methods, and classic tools remain available, while contextual orchestration by Gemini is gradually integrated into premium plans.
For Google, the challenge is to demonstrate that this layer of AI provides enough daily value to justify a recurring subscription.
Workspace moves closer to an intention-driven environment
Gmail, Docs, and Keep have historically been designed as three distinct applications, each serving a specific task. AI is beginning to blur this separation. When Gemini can retrieve information from Gmail, link it to a Drive file, and then prepare text in Docs based on a simple conversation, the specific application used becomes less important than the user’s expressed goal.
This is a profound shift in software design.
For decades, users had to learn where a function was located and how to use it. Conversational assistants now attempt to reverse this relationship: users describe the desired outcome, and the system selects the necessary functions.
Success will naturally depend on the accuracy of interpretation. An assistant that misinterprets a request or retrieves the wrong context could quickly create more work than it saves.
Google progressively builds its universal interface around Gemini
With Gmail Live, Docs Live, and Keep Live, Google showcases where it wants to take Workspace. Gemini is no longer just a sidebar offering to summarize or rewrite content. It is gradually becoming an interface positioned above the applications, capable of receiving a request and navigating across multiple services to execute it.
Voice accelerates this transformation because it makes interaction even more direct. There’s no longer any need to formulate precise queries or navigate through multiple menus: theoretically, you just need to explain what you want to achieve.
If Google succeeds in making this experience reliable, the change could be deeper than just a new generation of AI features. Gmail, Docs, and Keep may continue to exist, but users might need to think less about the application in which to start.
Ultimately, the true interface for Workspace could simply become the conversation with Gemini.


