OpenAI has formally launched ChatGPT Work, an enterprise-oriented capability that turns ChatGPT into an agent designed to act across users' apps and files, remain on a project for hours when necessary, and convert a goal into finished work. The rollout began on July 9, 2026, beginning on web and mobile, with staged availability across ChatGPT plans.
The central change is not simply another chat interface. ChatGPT Work is positioned as a system for completing multi-step work while operating within organizational controls. According to OpenAI's announcement of ChatGPT Work, the capability is powered by the GPT-5.6 frontier model and can help teams turn research, customer insights, and planning into outputs including marketing briefs, dashboards, and presentations.
For enterprise buyers, the launch puts the focus on a practical question: how can an AI agent access useful business context and tools without receiving unrestricted authority? OpenAI's answer combines connected workflows with controls over accessible context, connected tools, and actions that need human approval. That makes Work a meaningful extension of ChatGPT Enterprise rather than a separate replacement for the enterprise product. ChatGPT Work is designed for work that extends beyond a single prompt and response. OpenAI describes an agent that can coordinate information across apps and files, carry a project forward over a longer period, and produce a finished result from an initial objective.
The examples matter because they describe an operational workflow rather than isolated content generation. A team could use customer research to develop campaign assets, then adapt those assets for different markets. It could also bring research, planning material, and performance information together into a briefing, dashboard, or presentation. The degree of autonomy and the specific connected tools remain subject to the configuration selected by the organization and user.
OpenAI also highlights scheduled tasks. These can monitor data sources, update documents, and surface changes automatically. For teams that repeatedly assemble status information or refresh planning documents, that model could shift ChatGPT from an on-demand assistant toward a monitored workflow participant.
The launch materials identify a broader set of supporting capabilities around Work, including Sites, Plugins, a built-in browser, Cloud and Desktop synchronization, and administrative governance. The supplied information does not detail the configuration or availability of each component individually, so organizations should assess them within their own workspace and plan settings rather than assume every option is immediately enabled.
ChatGPT Enterprise remains the organization-level offering for deploying ChatGPT with enterprise privacy, security, and administration. ChatGPT Work is the agentic work capability being rolled out within that environment, adding a way to execute longer, connected tasks under defined controls.
| Area | ChatGPT Enterprise context | ChatGPT Work capability |
|---|---|---|
| Primary role | Enterprise workspace with privacy, security, and administrative controls | Agentic work that can turn goals into finished multi-step outputs |
| Work scope | Managed access to ChatGPT for organizational users | Can operate across apps and files and stay with a project for hours if needed |
| Governance | Admin settings, access management, spend controls, and enterprise-grade protections | Controls over accessible company context, connected tools, and actions requiring approval |
| Availability | ChatGPT Enterprise workspace | Rolling out incrementally, including eligible Enterprise and Edu workspaces |
The distinction is important for procurement and governance teams. An enterprise ChatGPT deployment can provide the managed environment, while Work introduces the operational question of what the agent is permitted to do inside it. The value of Work will therefore depend on policy design as much as model quality.
OpenAI says Work began rolling out on web and mobile on July 9, 2026. The company also describes desktop support for Chat, Work, and Codex across plans. Availability is incremental: Pro, Enterprise, and Edu are included in the initial expansion, followed by Plus and Business in the following days.
For enterprise access, OpenAI's Help Center states that Work is available to eligible ChatGPT Enterprise and Edu workspaces with Enterprise Key Management (EKM) enabled. It also identifies access management, browser and network controls, and spend oversight as enterprise controls relevant to Work across web, mobile, and desktop. That staged approach means enterprises should treat the announcement as the start of deployment, not a guarantee that every employee, platform, or workspace configuration is active at once. Administrators need to confirm workspace eligibility, enabled controls, and the intended scope of connected tools before introducing Work into sensitive processes.
Agentic systems can be more useful than standard chat assistants because they connect steps across information sources and tools. They also require more deliberate boundaries. OpenAI says teams retain control over three critical areas:
These controls are consequential when Work is used for customer research, planning, marketing operations, or recurring reporting. A useful implementation should define which information is appropriate for the agent, which actions can be automated, and where human review is mandatory. Spend controls add another layer of oversight for organizations managing agent use at scale.
Organizations evaluating connected AI workflows can work with Scalevise on AI architecture, workflow automation, and implementation planning that aligns agent capabilities with practical governance requirements.
OpenAI's announcement establishes the product direction: GPT-5.6-powered agents that can work across enterprise context and tools over longer task durations. The next practical signals will be how quickly plan-level access expands, how organizations configure approvals and connections, and which recurring workflows prove reliable enough to move from experimentation into routine operations.
The supplied launch information does not provide separate ChatGPT Work pricing. Buyers should therefore evaluate availability and commercial terms through their existing plan relationship and OpenAI's current workspace information, rather than treating Work as a separately priced product based on this announcement.
What is ChatGPT Work?
ChatGPT Work is an OpenAI agent capability that can act across users' apps and files, stay with a project for extended work, and turn a goal into finished outputs such as briefs, dashboards, and presentations.
When did OpenAI launch ChatGPT Work?
OpenAI publicly announced the rollout on July 9, 2026. It began on web and mobile, with availability expanding incrementally across plans.
Is ChatGPT Work separate from ChatGPT Enterprise?
No. ChatGPT Enterprise is the managed enterprise environment, while ChatGPT Work is the agentic capability rolling out for eligible Enterprise and Edu workspaces and other plans in stages.
What controls do enterprises have over ChatGPT Work?
Organizations can control the company context Work can access, the tools it can connect to, and when actions require human approval. OpenAI also identifies admin settings, access management, browser and network controls, and spend oversight.
Does OpenAI list separate pricing for ChatGPT Work?
The supplied launch material does not state separate pricing for ChatGPT Work. It describes staged availability across existing plans, including Enterprise and Edu.
ChatGPT Work marks OpenAI's move toward longer-running, connected AI work inside managed business environments. Its enterprise significance lies in the combination of GPT-5.6-powered agent workflows and explicit controls over context, tools, approvals, and spending. As the rollout expands, successful adoption will depend on whether organizations can apply those controls to useful, repeatable work without weakening their governance standards.