OpenAI presented "ChatGPT Work", which it defined as its biggest update for work within ChatGPT. The central novelty is an agent capable of taking an objective, searching for information in the connected apps, and sustaining complex projects for hours, dividing them into steps and completing them on its own. Here we tell you what changes concretely for marketing, product, and operations teams.

The announcement, made in a livestream in early July, is OpenAI's most direct move against Microsoft Copilot and marks a change of approach: from an assistant that answers questions to an agent that executes work end to end and delivers finished results. According to the coverage from TestingCatalog, the feature started rolling out first on the paid plans.

What ChatGPT Work is

ChatGPT Work is an agent designed for long and complex tasks, not for quick answers. You give it an expected result and it researches, works on files and connected apps, and produces finished deliverables: documents, spreadsheets, presentations, reports, and sites. While it works, you can follow its progress, answer its questions, change course, and approve important actions before it executes them.

Scheduled tasks

One of the most relevant capabilities is keeping projects in motion with scheduled tasks that run once, repeat on a schedule or a trigger, or monitor changes. This brings it closer to the idea of a collaborator that works even when you're not in front of the screen.

How it's accessed and on which platforms

The rollout started on web and mobile for the Pro, Enterprise, and Edu plans, with access for Plus and Business planned shortly after. In parallel, OpenAI made globally available a desktop app for Windows and Mac that integrates Chat, Work, and Codex in one place, even for the free plan. The logic is clear: unify the experience and put the agent into the daily workflow.

What changes for each team

Marketing

Market research, campaign drafts, content calendars, and reports that previously took hours can be delegated as a project and received already put together to review. The professional's role shifts toward strategy and critical editing.

Product

Feedback analysis, documentation, competitor research, and site prototypes or specs can be generated more autonomously, freeing up time for product decisions.

Operations

The combination of multi-step tasks and scheduled tasks makes it possible to automate follow-ups, data consolidation, and recurring reports that today consume manual time.

This move is framed within the parallel launch of new models, which we cover in the article about GPT-5.6 and the Sol, Terra, and Luna family.

What to keep in mind before adopting it

An agent that acts on its own demands more judgment, not less: reviewing the deliverables, controlling which actions it approves, and taking care of what data is connected is indispensable. Human supervision remains the factor that separates a safe adoption from a risky one, something consulting firms like McKinsey in its report on the state of AI agree on.

Recommended Coderhouse courses

• AI Agents Course: to understand in depth how autonomous agents like ChatGPT Work work and how to take advantage of them.

• AI Automation Course: ideal for operations teams that want to automate processes with AI.

• Introduction to Artificial Intelligence Course: perfect for non-technical profiles who want to incorporate these tools with judgment.

Frequently asked questions

What is ChatGPT Work?

It's an agent within ChatGPT that takes an objective, works on connected apps and files, and executes complex multi-step tasks for hours, delivering finished results like documents, spreadsheets, and presentations.

How does it differ from traditional ChatGPT?

The classic ChatGPT answers questions conversationally. ChatGPT Work is designed to execute long projects more autonomously, following its progress and approving important actions along the way.

Who can use it?

The rollout started on web and mobile for the Pro, Enterprise, and Edu plans, with Plus and Business joining shortly after. Availability may vary according to the region and the plan.

Does it replace teams' work?

More than replacing, it shifts the human focus toward strategy, supervision, and critical editing. Reviewing the deliverables and controlling the agent's actions remain indispensable tasks for people.