ChatGPT Mobile Adds Voice-Based Agentic Features
OpenAI is bringing voice-controlled AI workflows to the ChatGPT mobile app. Announced on September 23, 2026, the update allows users to initiate work tasks through spoken instructions. Consequently, users can create documents, draft emails, and summarise messages while using their phones.
The rollout extends ChatGPT’s agentic capabilities beyond desktop and web experiences. Moreover, it allows users to move between voice conversations and written interactions within the app.
Voice Commands Expand ChatGPT Work
The update brings voice functionality to ChatGPT Work on mobile. Eligible subscribers can now speak instructions to complete professional tasks through the Work interface.
For example, users can request a document, draft an email, or summarise Slack messages. Additionally, they can create presentations, build websites, and use the cloud browser. These capabilities allow ChatGPT to handle more complex workflows through natural-language commands.
Furthermore, users can access connected apps and supported tools through voice. ChatGPT can use the permissions and integrations available to their accounts. However, existing access restrictions and approval requirements still apply.
OpenAI’s September 23 release notes confirm that Voice now supports plugins on web, iOS, and Android. Therefore, users can interact with supported connected applications during voice conversations and follow written responses in the chat.
Subscription Access and Availability
The new features offer different levels of access depending on the ChatGPT plan.
- Plus and Pro: Users can access voice-based workflows in ChatGPT Work, depending on their plan’s available features.
- Free and Go: Users can use voice conversations with the plugins and connected apps supported by their plans.
However, voice access within Work requires both Voice and Work permissions. Existing usage limits and workspace controls also remain in effect.
Moving Between Voice and Text
Another change concerns continuity between spoken and written interactions. Users can start a task while away from their computers and continue working on it later.
For instance, a user can begin drafting a document through voice on a phone. Subsequently, they can resume the conversation on a desktop and review the written output.
Moreover, unfinished tasks in Work can continue in text after users end a voice session. This allows users to switch interaction methods without restarting their work.
The update builds on OpenAI’s July 2026 introduction of GPT-Live. That release brought more natural voice conversations, including real-time responses and interruptions. Now, the September expansion connects those conversational capabilities with professional workflows on mobile devices.
As a result, ChatGPT is extending voice beyond conversational assistance into document creation, connected applications, and task execution. Availability and functionality will continue to depend on account eligibility, permissions, and supported integrations.
