OpenAI turns ChatGPT into a shared work space
OpenAI launched ChatGPT Space on 29 September as a shared home for documents, files and work with AI agents. Teams can co-edit Pages, bring context into one space and assign work to ChatGPT, Codex or a Dot. Space is available to Pro, Business and Enterprise users on web and desktop; mobile users can find, read and share pages, with creation and editing to follow. The product replaces the earlier Library experience for these plans. OpenAI is trying to move ChatGPT into the persistent documents and team context where work is completed, giving its models a continuing place in an organization's daily process rather than a series of isolated conversations.
Pages are what OpenAI calls ‘living documents’ that can include text, research, charts, prototypes and connected information. Teammates can make changes in real time, comment and ask an agent to revise a page or handle a follow-up. A page can carry instructions about what sources to consult and when to update, allowing recurring work to live with its output. The shared Space can hold files, slides, spreadsheets and other material; members can return to the same project context instead of copying a prompt and attachments between chats. OpenAI says each editor works with their own ChatGPT when collaborating on a page, so the collaboration is both human and model-mediated.
The launch extends into existing channels. Business and Enterprise users can mention ChatGPT in Slack and Microsoft Teams, where colleagues can add context and refine a result in a shared conversation without every participant requiring an individual ChatGPT license. A Meetings plugin in beta on the macOS desktop app for Pro and Business users creates private notes and suggested actions that can be brought into Space. OpenAI says the recorded audio is deleted once notes are ready. Collaborative slide editing is planned for the coming weeks, while spreadsheet collaboration and mobile editing are still to come. Those staged releases distinguish the working Pages product from the broader suite OpenAI is describing.
OpenAI's Space page says a team can set project priorities and invite members or a group under the workspace's sharing rules. View and edit rights remain configurable, a necessary detail when a page mixes employee contributions with material drawn from connected tools. An automation can refresh a document according to specified sources and timing, turning a static report into a maintained artifact. This could be useful for operating dashboards, research briefs and project plans, but it also makes permissions and source provenance central to adoption. A team will need to trust who can see and change shared context before it asks an agent to act on that context.
The timing matters because Dots, the new persistent agent, are designed to work within this same environment. Space gives a Dot shared knowledge and visible output; Dots give Space a worker that can carry assignments forward. OpenAI's product page describes Library as becoming Space rather than merely adding a collaborative feature to the old file store. The company is also opening more of ChatGPT to third-party plugins, so a page may increasingly become a surface for work that begins in another service. For an enterprise customer, the potential value is fewer handoffs among chat, documents, meetings and project tools. The competitive question is whether the integrated workflow is good enough to displace familiar collaboration habits.
Analysis
Space can convert occasional ChatGPT use into organizational dependence if shared documents, instructions and agent outputs accumulate around recurring projects. That creates a stronger retention mechanism than an individual's prompt history and gives OpenAI more surfaces to sell premium seats and usage. The Slack and Teams option widens reach even where a company keeps its incumbent collaboration system, while the lack of a license requirement for every participant can speed a trial. The commercial tension is that OpenAI must prove its Pages and agent work exceed the value of existing suites and connectors; the cost of migration is social as much as technical when teams have established review and sharing habits.