ChatGPT Voice Adds Multistep Agent Work to Mobile

OpenAI is adding its most advanced voice system to ChatGPT’s website and mobile app, enabling spoken delegation of single or recurring tasks. [2] Those tasks can span multiple steps, use a browser, and involve disparate apps. [2]

On mobile, OpenAI says voice can trigger workflows such as drafting documents and summarizing emails. [1] The rollout brings capabilities previously available through GPT-Live in ChatGPT’s desktop Work and Codex tabs to the mobile app. [1]

Plus and Pro users can use the phone’s Work tab to create documents, draft emails, summarize Slack messages, build sites, create presentations, use the cloud browser, and access areas including finances. [1] Free and Go users will instead get access to plugins and connected apps. [1]

OpenAI also says voice conversations will produce richer text output, allow easy switching between text and voice, and support starting on mobile before resuming on desktop. [1] Anthropic recently eased mobile-to-desktop handoffs and merged Claude’s Cowork and Chat interfaces, while OpenAI continues to keep chat and workspaces separate. [1]

Voice is becoming an execution interface rather than only a conversational one: users can delegate multistep, browser-assisted work and then move between mobile and desktop. Access remains tiered, with paid users receiving broader creation and cloud-browser capabilities while Free and Go users are limited to plugins and connected apps. The clearest competitive contrast is interface design: Anthropic has merged Cowork and Chat, whereas OpenAI still separates chat from workspaces despite supporting cross-device continuation. What the sources don't address: what safeguards govern voice-triggered multistep actions across browsers and connected apps.