OpenAI expands ChatGPT with voice agent for all users, intensifying competition for autonomous assistants

New voice features now available on web, mobile, and CarPlay

On Wednesday, 23rd, OpenAI rolled out ChatGPT’s voice agent capability to all users, both on its web platform and mobile app. The feature lets users issue spoken commands to perform tasks like creating documents, searching the internet, and drafting emails. Responses appear on screen as formatted text, and users can switch between speaking and typing or resume conversations later. The functionality is also available on Apple’s CarPlay system for vehicles.

Subscription plan differences determine functionality

Users on the Plus and Pro plans have access to a broader range of voice actions in ChatGPT Work. They can generate documents, presentations, spreadsheets, images, and web pages using ChatGPT Sites, as well as use a cloud browser and integrate data from services like email, Slack, and Google Drive to start new projects. Free or Go plan subscribers can use voice commands with the tools and integrations already available in their standard conversations, though OpenAI has not yet specified whether usage limits will apply to these plans.

From chatbot to autonomous agent: a strategic shift in the industry

The launch reflects a trend in the artificial intelligence market, which is moving from assistants that simply answer questions to agents capable of acting on behalf of users. Voice is seen as the ideal medium for deploying these agents in situations where hands or attention are occupied. The same logic drives recent updates to Apple’s Siri, Google’s Gemini Live, and Meta’s Muse, a personal assistant launched this month.

Computational impact and market projections

Autonomous agents require significantly more processing power than traditional chatbots. According to consulting firm Gartner, these systems consume five to thirty times more processing units (tokens) per task, resulting in higher operational costs. Gartner predicts that an average Fortune 500 company will increase its number of AI agents from fewer than 15 in 2025 to over 150,000 in 2028, signaling a rapid expansion in corporate adoption.

Use cases and uncertainties about limits and languages

OpenAI suggests that the voice agent can be useful for summarizing priorities between meetings, planning trips, getting shopping recommendations, and receiving financial advice. The company did not specify whether the feature will have usage restrictions on free plans or whether it will function the same way in Portuguese.