OpenAI has integrated business agents into ChatGPT that can perform tasks without human involvement

OpenAI launches new cloud “work agents”

OpenAI announced the launch of Business, Enterprise, Edu, and Teachers subscription plans with access to new cloud AI agents. These assistants, as *The Verge* reports, are integrated into ChatGPT and can act as autonomous employees within a corporate environment.

What work agents can do
- Automate routine tasks—from gathering product feedback worldwide to drafting emails in Gmail.
- Example from practice—a bot independently searches for the needed information, compiles a report, and sends it to Slack. The process works like this: the agent gathers context from specified sources while adhering to internal team guidelines, then requests approval before executing key actions.

How it works for companies
- A single set of agents—companies create an agent once and share it in a common workspace.
- Continuous improvement—the shared use gradually enhances the agents’ skills and efficiency.

Why this step now
OpenAI has been developing this system to keep pace with competition from other AI‑agent providers. After the success of the OpenClaw project (which involved Peter Steinberger, now an OpenAI employee), competition intensified. Anthropic already offers similar tools for working with user files and a separate platform of autonomous agents.

What it means for GPT models
Deploying work agents may signal the end of the era of custom GPT chatbots announced in 2023. OpenAI positions the new assistants as an “evolution” of GPT: GPT models will remain available while teams test work agents on their enterprises. In the near future, the company plans to simplify the transition from older GPT models to new work AI agents.

Thus, OpenAI is introducing more powerful and flexible tools for corporate users while simultaneously changing the approach to personalized chatbots.