OpenAI brings workspace administration into ChatGPT Work and Codex
OpenAI's Admin plugin lets ChatGPT Work and Codex workspace administrators inspect usage, manage access, and complete supported changes without leaving a conversation.
OpenAI has introduced an Admin plugin for ChatGPT Work and Codex that turns routine workspace administration into permission-aware conversational workflows. Administrators can inspect activity and credit usage, manage members and groups, control access, adjust usage limits, and review spending requests from one interface.
What the Admin plugin changes for workspace teams
The plugin connects questions, investigation, and supported actions in the same conversation. An admin can ask which teams are nearing credit limits, inspect effective permissions, update a group, or review a spending request, then receive a structured result showing what happened. AI Industry Today independently reported the same August 25 launch and the same core administration scope.
This is a workflow change rather than a new permission layer. OpenAI says the plugin operates within the administrator's existing role, workspace policies, approval requirements, and supported controls. It does not grant broader access simply because a request is phrased conversationally.
Recurring checks can route decisions to Slack or Teams
OpenAI also describes recurring administrative automations. The plugin can monitor feature-access requests, grant access when predefined criteria are met, and route exceptions for review. Pending usage requests can be sent to Slack or Microsoft Teams, where authorized reviewers approve or deny them.
Those flows matter because they keep the decision boundary visible. The system can perform a supported action, but admins still need to define the policy, approve exceptions, and verify the resulting change. Teams should treat the conversational layer as an interface to existing controls, not as a replacement for role design or change management.
OpenAI's own deployment shows the intended pattern
OpenAI says its IT team uses the plugin to combine workspace data, reports, and operational actions. The company reports that a ChatGPT Work agent handling employee IT requests resolved about 45% of ticket volume at the time of reporting, while support volume roughly doubled. Those figures are OpenAI's internal account, not an independently audited benchmark.
The practical pattern is clear: connect usage data to the next permitted action, keep approvals for broader changes, and return an auditable result. That can shorten the path from “who is blocked?” to “what policy-compliant change is needed?” without hiding the responsibility of the administrator.
Availability and controls
OpenAI's installation instructions say to enable the plugin in workspace settings and install it from the Plugins directory in ChatGPT Work on the web or desktop app. The announcement targets workspace administrators; readers should confirm that their plan, workspace policy, and regional rollout expose the plugin before planning a migration.
The next operational check is governance: review which roles can install the plugin, which actions require approval, where Slack or Teams requests are routed, and how changes are logged. Conversational administration is useful only when those boundaries remain explicit.
Sources
Try the related loot
Put six hosted Workers AI models behind Cloudflare AI Search