ChatGPT Work | Agent Creates Slides, Documents, and Sites

OpenAI has introduced ChatGPT Work, an agent designed to handle longer projects across connected apps, files, websites, and desktop tools. It can gather project context, create editable slides, documents, spreadsheets, and sites, then continue refining the work while the user reviews progress and controls important actions.


ChatGPT Work creating slides, documents, spreadsheets, and interactive sites

{getToc} $title={Table of Contents}

ChatGPT Work moves from individual prompts to complete projects


ChatGPT Work is built for tasks that require several connected steps instead of one immediate response. It can break a goal into smaller actions, gather information from relevant tools, produce working materials, and stay with the project as it develops. Users can follow its progress, answer questions, redirect the workflow, and approve actions that require additional control.


For designers and content teams, the important change is continuity. A project can begin with research or source files, continue into a campaign brief or structured document, and then become a presentation, spreadsheet, marketing asset, or interactive site without losing the context carried through the earlier steps.



Connected apps provide the context behind each deliverable


ChatGPT Work can connect to tools such as Google Drive, SharePoint, Slack, Microsoft Teams, email, calendars, project trackers, and customer-management systems through plugins. The agent can use that material to understand the task and build an output that reflects the information already used by the team.


This can reduce the manual work of collecting references before beginning a creative project. A campaign brief, for example, may pull from customer research, internal discussions, project documents, and existing performance information. The same context can then support the creation of presentation slides, visual assets, regional variations, or other project materials.


Slides, documents, spreadsheets, and Sites remain editable


The agent can create structured materials such as presentations, reports, spreadsheets, project trackers, and launch plans. These outputs are intended to be working files rather than static answers, allowing teams to continue reviewing, editing, and updating them as the underlying information changes.


OpenAI is also introducing Sites in ChatGPT as a public beta. Sites can turn project information into shareable dashboards, prototypes, launch calendars, internal portals, trackers, and interactive reports. The result can be tested inside ChatGPT and shared through a URL when the project needs a more visual or interactive format.


Desktop tools and scheduled tasks extend the workflow


The updated ChatGPT desktop app can work with local files, desktop applications, and a built-in browser. This allows ChatGPT Work to gather information from websites, open supported files, interact with tools, and carry out steps that would normally require moving manually between several windows.


Scheduled Tasks can also repeat actions or respond when something changes. A team could ask ChatGPT to monitor feedback, refresh a recurring report, update a presentation when new information arrives, or review project activity before a regular meeting. This makes the agent relevant for ongoing production workflows, not only one-time document creation.


IMPORTANT: ChatGPT Work is rolling out on web and mobile for Pro, Enterprise, and Edu plans, followed by Plus and Business. The updated desktop app is available on Mac and Windows, but usage, connected tools, approvals, and organizational controls depend on the selected plan and admin settings.{alertWarning}

Daisuki's Take: What This Means for Designers


ChatGPT Work is relevant because creative projects rarely begin and end with a single prompt. Designers often need to gather research, organize feedback, prepare a brief, create assets, and adapt the result for different formats. Keeping those stages connected can reduce the repeated setup between tools.


The creation of slides, documents, spreadsheets, and Sites also makes the agent more useful for presenting design work. A visual idea may need supporting research, a launch plan, a dashboard, or an interactive prototype before other teams can review and approve it.


We would still keep human review at every major creative decision. Connected context can improve the output, but designers need to verify source accuracy, visual hierarchy, brand consistency, permissions, and whether the final material communicates the project clearly.



Sources and Recommended Links