OpenAI has announced a new feature called “workspace agents,” signaling a major evolution in how businesses can use ChatGPT. The update moves the platform beyond simple conversations, enabling it to function as an autonomous assistant capable of handling ongoing workplace tasks without constant human direction.
Powered by Codex for Smarter Task Execution
At the core of this new feature is Codex, which powers the agents’ ability to execute complex, multi-step workflows. Unlike earlier custom GPT tools, workspace agents are designed to persist over time, maintaining context across projects and integrating with external applications. This allows them to perform tasks such as drafting reports, writing code, and responding to workplace messages with minimal supervision.
One of the defining characteristics of workspace agents is their ability to operate continuously. Because they are cloud-based, these agents can keep running even when users are offline. This ensures that business processes—such as data updates, communications, and scheduled tasks—continue without interruption, improving overall efficiency within organizations.
Built for Team Collaboration
OpenAI is positioning workspace agents as collaborative tools rather than individual assistants. Teams can design a single agent tailored to a specific workflow and share it across departments. Over time, these agents can be refined collectively, allowing organizations to improve performance and consistency as more people contribute to their development and usage.
Creating a workspace agent is designed to be user-friendly. Through a dedicated section in ChatGPT, users can describe the workflow they want to automate. The system then helps outline the process, connect relevant tools, and test the agent before deployment. Once active, agents can operate on schedules or respond to triggers, such as incoming data or specific business events.
Supporting Complex Organizational Workflows
According to OpenAI, the feature is particularly useful for workflows that require coordination across teams. Workspace agents can gather information from connected systems, follow established procedures, and even request approvals when necessary. This makes them suitable for handling tasks that involve multiple stakeholders and decision points, helping organizations maintain smoother operations.
Rising Competition in Agentic AI
The launch comes at a time when the race to develop autonomous AI systems is accelerating. Major technology players such as Google, Microsoft, and Amazon are investing heavily in similar technologies. This growing competition reflects a broader industry shift toward AI tools that can independently complete tasks with limited human oversight.
Security Measures and Safeguards
OpenAI acknowledged potential risks associated with autonomous systems, including prompt injection and data security concerns. To address these issues, organizations can set strict controls on what data agents can access, require human approval for sensitive actions, and monitor for malicious inputs. These safeguards are intended to ensure that automation does not come at the expense of security.
Workspace agents are currently available in a research preview for ChatGPT Business, Enterprise, Edu, and Teachers plans. OpenAI stated that the feature will remain free to use until May 6, 2026. After that date, it will transition to a credit-based pricing model.
Despite the introduction of workspace agents, OpenAI confirmed that existing GPT tools will continue to be supported. The company also plans to make it easy for users to convert their current GPTs into workspace agents, allowing for a smooth transition and continued use of previously built solutions.
