OpenAI announced on June 2, 2026, a series of new features for its intelligent agent tool Codex, aimed at enterprise white-collar workers. The goal is to expand its application scenarios from software engineering to broader knowledge work fields, accelerating its competition in the enterprise market.

According to an internal report released by the company, the current weekly active users of Codex have exceeded 5 million, growing more than six times since the desktop application was launched in February of that year; among them, knowledge workers account for about 20% of the total users, and their growth rate has exceeded three times that of developers.

QQ20260603-092300.jpg

To deeply attract this group, OpenAI has introduced ready-to-use plugin toolkits for six specific positions: data analysis, creative production, sales, product design, equity investment, and investment banking. By integrating dedicated instructions and context information, Codex can now highly resemble independent completion of specific tasks.

Another major upgrade in terms of technology and functionality is the introduction of the "Sites" feature, which allows Codex to directly output work results as hosted interactive websites rather than traditional local files. For this, OpenAI has partnered with ecosystem partners such as Wix, Base44, Replit, Lovable, Figma, and Emergent. Additionally, the new annotation feature allows users to select specific parts of a file to achieve more precise contextual commands.

OpenAI,人工智能,AI

This move follows its strategic action three weeks earlier, when it invested over $4 billion to establish a joint venture - OpenAI Deployment Company, aiming to deeply embed AI into global corporate workflows. In the context of competitors like Anthropic pushing forward financial-grade intelligent agent projects, OpenAI is shifting from an early consumer focus to a high-intensity B-end penetration. As AI systems take on more meaningful corporate back-end tasks, overcoming infrastructure and workflow integration challenges will become the decisive factor in the next phase of intelligent agent ecosystem competition.