OpenAI Launches ChatGPT Work: Your AI Agent Now Does Complex Tasks

OpenAI launched ChatGPT Work on Thursday, an AI agent integrated into its chatbot to serve as an autonomous work platform. Powered by GPT-5.6, this agent can execute complex, multi-step tasks across users' email, calendars, and code repositories. It aims to move ChatGPT beyond question-and-answer interactions to actively producing finished documents, spreadsheets, and presentations. This marks a significant shift for OpenAI, repositioning ChatGPT as a workplace tool amidst its IPO preparations.

The core architecture of ChatGPT Work features a persistent, cloud-based virtual machine, ensuring availability across devices. This marks a departure from competitors requiring local machine power. The product leverages MCP-based plugins to connect with services like Gmail and GitHub. Initially rolling out to Pro, Enterprise, and Edu users, it will soon be available to Plus and Business users, underscoring OpenAI's mission of broad accessibility.

ChatGPT Work competes directly with Anthropic's Claude Cowork and Microsoft's Copilot. All three products envision persistent cloud-based AI agents for task execution. OpenAI's advantage lies in its vast user base, with over 900 million weekly active users, aiming for faster adoption by offering ChatGPT Work to $20/month Plus subscribers. The company is generating $2 billion monthly, with enterprise revenue nearing parity with consumer revenue.

Data privacy is managed through user control and OpenAI's existing enterprise security. Users can opt out of having conversations improve future models, aligning with assurances that business data is not used for training. This accessibility and autonomous capability aim to prove enterprise AI can generate substantial revenue, a critical narrative for OpenAI's IPO. The agent can perform analytical tasks, reducing project timelines from months to weeks.