The Pulse
OpenAI Brings Voice-Controlled Workflows to ChatGPT Mobile
OpenAI is bringing voice-based agentic workflows to the ChatGPT mobile app, allowing Plus and Pro subscribers to create documents, draft emails, summarize Slack messages and build presentations by speaking. Free and Go users will be able to
AI.info Team ·
150 Million Weekly Voice Users Get a More Capable ChatGPT
More than 150 million people use ChatGPT’s Voice and Dictation features each week, according to OpenAI. On September 23, the company began extending voice from conversation into mobile work tasks, allowing users to ask ChatGPT to create documents, draft emails and summarize Slack messages through the phone app.
The rollout brings ChatGPT Work into mobile voice conversations for Plus and Pro subscribers. Users can also ask the system to build sites, create presentations, operate the cloud browser and work with areas such as Finances inside ChatGPT. OpenAI’s announcement was first reported by TechCrunch.
The change matters because the phone version of ChatGPT previously offered a relatively narrow voice experience: users could talk to the assistant, but the voice interface was not designed to initiate complex work on the device. OpenAI now connects that conversational layer to the longer-running workflows already available through ChatGPT Work.
Plus and Pro Get the Full Work Tab
Paid subscribers can use the Work tab on mobile to hand off multi-step tasks by voice. A spoken request might ask ChatGPT to turn source material into a document, prepare a presentation, summarize a Slack conversation or draft an email. The user can then continue the exchange by voice or inspect the written result in the app.
OpenAI also says Plus users can move between text and voice during a conversation. That matters for tasks where a user wants to dictate the initial instruction but review or edit the output visually. A conversation started while away from a desk can also be resumed on a desktop computer, according to the company’s announcement.
Free and Go users receive a narrower version of the update. They can use voice with supported plugins and connected apps, but TechCrunch reports that the full Work tab remains limited to Plus and Pro subscribers.
GPT-Live Connects Conversation With Deeper Work
The mobile rollout builds on GPT-Live, OpenAI’s voice model family introduced in July. OpenAI describes GPT-Live as a full-duplex system that can listen and speak at the same time, while delegating search, reasoning and more involved tasks to other models in the background. The company says the architecture lets the voice session continue while deeper work takes place.
OpenAI’s GPT-Live announcement says the model can decide whether to keep listening, pause, interrupt or invoke a tool. The company also reported that more than 150 million people use Voice and Dictation each week, although that figure covers both features rather than voice conversations alone.
ChatGPT Voice gained access to stronger reasoning models earlier in September. OpenAI’s release notes say Voice can use GPT-5.6 or GPT-6 Astra when a request requires search or deeper reasoning, with availability and usage limits depending on the user’s plan.
Mobile Extends a Rollout That Started on Desktop
OpenAI first connected voice to ChatGPT Work and Codex in the desktop app in July. That update allowed users to issue complex commands, check on agent progress and coordinate multiple agents through spoken instructions. In one example covered by TechCrunch, a user asked ChatGPT to create a discussion thread, make a pull request and investigate a software bug with one command.
The desktop launch also showed the boundary between voice conversation and computer control. ChatGPT could use computer-use capabilities to inspect websites and applications, while Codex could continue work remotely from an iOS device. The September 23 update shifts more of the workflow itself onto the phone rather than using the phone only as a remote control.
OpenAI is not combining Chat and Work into one interface. TechCrunch reports that the company is keeping the two areas separate, even as voice can now reach Work from mobile. That distinction gives users a dedicated place for longer-running tasks while preserving ChatGPT’s conventional conversation mode for quick questions and answers.
The Phone Becomes the Starting Point, Not the Whole Workspace
Voice is useful for starting work when typing is inconvenient, but the new feature still depends on a written interface for checking results and handling details. Drafting an email or asking for a summary can begin hands-free; approving an action, correcting a name or reviewing a presentation may still require careful visual inspection.
The mobile handoff also raises practical questions about permissions. Work can use connected apps and files only within the access granted to the account, and tasks that involve sensitive actions may require confirmation. OpenAI’s broader ChatGPT Work materials describe a system that can operate across connected apps and files while allowing users to follow progress, change direction and approve important actions.
For now, the concrete shift is straightforward: Plus and Pro users can start ChatGPT Work tasks with their voice on a phone, while Free and Go users gain voice access to supported plugins and connected apps. The next test is whether people use mobile voice to begin substantial work often enough to justify the additional permissions, subscription limits and review steps that follow.