OpenAI brings team agents to ChatGPT

OpenAI for ChatGPT Workspace Agents announced the new feature called . With this feature, the company introduces shared agents that teams use jointly, undertake multi-step tasks and maintain workflows. Moreover, these agents do not just respond to a single command and stand aside, they break the mission into parts, update the outputs and continue to work with affiliated services. Like this ChatGPTgoes beyond the writing and research side and dives more directly into team work.

However, OpenAI positions this system as a further evolution of the existing proprietary GPT approach. The main difference here is that the logic of responding within a single chat is abandoned and a work flow is established around files, tools, code and corporate permissions. According to the framework provided by the company Workspace Agentsmeaning workspace agents carry task flows that teams use over and over again. For this reason, the product finds its place in the joint processes of teams rather than individual use.

On the other hand, according to OpenAI, these agents Codex-based agents running in the cloud serves as. When given a task, it not only produces descriptions, it divides the job into steps, writes or runs code if necessary, pulls data from connected applications and prepares specific outputs. Additionally, OpenAI clearly conveys that these agents access a workspace with files, tools, and memory. This way, the task doesn’t just stay in the chat window, but spreads across a broader workflow.

Moreover, the highlight here is that workspace agents can be shared within the team. A company or team prepares an agent once, then prepares that agent both ChatGPT in both slack can be used jointly. OpenAI also states that agents can be shared in the workspace directory, deployed via connection, and tracked through analytics tools along with version history. These details move ChatGPT away from a personal chat tool and closer to a business assistant shared by teams.

OpenAI agents move team work to the cloud

However, this innovation is not limited to the sharing side. According to the information provided by OpenAI, workspace agents, Google Drive, Google Calendar, slack And SharePoint It can connect to services such as. In addition, teams are special MCP servers and can also add additional tools. The framework in the help documentation shows that these agents can be tested and published, receive scheduled runs, and take on tasks in Slack channels. Thus, tasks such as preparing regular reports, collecting content, answering internal queries and following specific workflows become more permanent.

In fact, the real change here is exceeding the “ask and get answers” limit. OpenAI says the agents run in the cloud and can continue the task even if the user is not present. In other words, the system is not only able to react to the command given at that moment, but also can continue the workflow. However, OpenAI also states that it plans to add deeper action support with more automatic triggers, performance dashboards and business tools in the future. This information shows that the product is usable in its current form but is still open to expansion.

In addition, OpenAI has not opened this feature to everyone. According to the company’s announcement Workspace Agentsat this stage ChatGPT Business, Enterprise, Edu And teachers It is included in the plans as a research preview. In other words, the feature is offered in corporate and education-oriented plans rather than individual free usage. This framework on the access side reveals OpenAI’s initial focus on teams and organizations.

Frankly, the picture presented by the company shows that ChatGPT will not remain just a tool that helps thinking. When teams set up a workspace agent, it works with files, runs code, connects to enterprise tools, updates output, and makes it public. Moreover, details such as link sharing, version history, analysis tools, and scheduled runs reveal that this use is not a one-off. Like this Workspace Agentsbecoming one of OpenAI’s most concrete steps that connects ChatGPT more tightly to team work.