Model: ChatGPT · Work
On 9 July 2026, OpenAI launched ChatGPT Work — an agent that researches, works across your connected apps and files, and produces finished documents, spreadsheets, presentations and reports. It also replaced the App Directory with a new Plugin Directory. Here’s what an agentic ChatGPT changes for whether your brand gets named.
Source: ChatGPT release notes — Introducing ChatGPT Work (9 July 2026)
- On 9 July 2026, OpenAI introduced ChatGPT Work, an agent that researches, works across connected apps and files, and builds finished documents, spreadsheets, presentations, reports and Sites.
- OpenAI also replaced the App Directory with a Plugin Directory, where plugins package skills, apps and app templates for specific workflows.
- Work is rolling out to paid plans except Free and Go, in all supported regions, with Pro, Pro Lite, Enterprise and Edu first.
- Its built-in browser gathers information from websites to build those deliverables — so which pages it reads decides which brands end up in the output.
- This is a new answer surface to watch: test how Work researches your category, not just how ChatGPT chats about it.
What OpenAI actually launched
On 9 July 2026, OpenAI introduced ChatGPT Work. In OpenAI’s own words, “ChatGPT Work is an agent for longer, more involved tasks.” According to the ChatGPT release notes, it can research and analyse information, work across connected apps and files, and create finished documents, spreadsheets, presentations, reports and Sites. You can follow its progress, answer its questions, change direction, and approve important actions as it works.
Work can also keep projects moving through Scheduled Tasks that run once, repeat on a schedule or trigger, or monitor for changes. It arrived alongside two other launches the same day: a new ChatGPT desktop app that brings Chat, Work and Codex together, and ChatGPT Sites for building lightweight web pages and apps. Together they mark a shift from ChatGPT as a place you ask questions to a place that does multi-step work on your behalf.
Why an agent is a new answer surface
A chat answer is read by one person and forgotten. An agent’s output is different: it becomes a report, a shortlist, a spreadsheet of vendors, a slide that gets circulated and acted on. When ChatGPT Work compiles “the five tools we should evaluate” or “a competitor overview,” the brands it names travel much further than a passing mention in a chat window.
That makes Work its own place to be visible, not just a wrapper around the same chat. The rules for who gets included are the ones we set out in how LLMs decide which brands to cite — but the stakes are higher, because the output is a deliverable someone relies on. If you are missing from the research an agent hands to a decision-maker, you are missing from the decision, a theme we keep returning to in is your brand ready for AI search?
What the Plugin Directory changes
Alongside Work, OpenAI is “replacing the App Directory with the Plugin Directory.” Existing app connections are unaffected, but plugins now package skills, apps and app templates for specific workflows, available from ChatGPT on web and desktop — including inside Work and Codex. In practice, the Plugin Directory is a new discovery layer: it shapes which tools ChatGPT can reach for when it completes a task.
For brands that offer a product an agent could call, that discovery layer is a visibility channel in its own right, much like the commercial placements we track in ChatGPT advertising. For everyone else, it is a reminder that being “named” by AI increasingly means being reachable by it — the same groundwork we describe in building a GEO strategy.
How ChatGPT Work gathers sources
The detail that matters most for AI presence is how Work does its research. On the new desktop app, Work can use local files and desktop apps with your permission, and its built-in browser lets ChatGPT gather information from websites and work with supported web-based tools and files. In other words, Work reads the live web to build its deliverables — so the pages it trusts feed directly into the documents it produces.
That ties agent visibility to the same discipline behind text answers: clear, current, well-structured content a model can read and reuse. It is the logic Google has been formalising in preferred sources for AI Overviews, and it compounds with memory — because a connected agent can lean on what it already knows about a user, much like the effect we explored in ChatGPT memory sources. If your category pages are thin or out of date, an agent researching your space will simply build around you.
What ChatGPT Work does on your behalf
ChatGPT Work, an agent for longer, more involved tasks, rolling out to paid plans except Free and Go from 9 July 2026.
Why deliverables raise the stakes
When an agent turns research into a finished report, it compresses a whole category into a handful of named options — and those names carry the weight of a recommendation. A vendor shortlist in a slide deck is not a casual mention; it is the basis for a shortlist someone acts on. The gap between being in that deck and being left out is the gap between being considered and being invisible.
Personalisation widens the gap further. Because Work draws on connected files and prior context, the same brief can produce different shortlists for different users — the same effect we flagged when ChatGPT began bypassing memory for fast answers. The practical consequence is that one spot check tells you little. What matters is the pattern across many agent runs: how often does your brand make the deliverable at all?
Where ChatGPT Work is rolling out
On web and mobile, Work is rolling out to paid plans except Free and Go. Pro, Pro Lite, Enterprise and Edu users get access first, with Plus and Business users following over the coming days, and it is available in all supported regions. Enterprise and Edu workspaces get a two-week preview period during which Work is off by default; admins can opt out before it turns on automatically at the end of the preview.
The reach is what makes this matter. This is a paid, work-focused audience — the people making buying decisions — using an agent that names brands inside the documents they ship. Region and rollout date still shape what any given user sees, as we covered in how geopolitical events reshape AI visibility, so coverage will vary as the rollout widens. Note too that public publishing for the sibling Sites feature is not available in the UK, EEA or Switzerland at launch, even though Work itself is.
What the launch means for brands
Every time ChatGPT gains a new way to answer, the answer can change with it. We have watched that play out release after release — from the retirement of GPT-4.5 to the GPT-5.5 Instant Mini fallback swap and the new voice model in ChatGPT Voice. An agent surface is a bigger shift than any single model swap, because it changes not just the wording of an answer but what the answer is — a finished piece of work rather than a reply.
If you have only ever checked how you appear in a typed ChatGPT chat, an agent that researches and builds is a blind spot worth opening. The web-search grounding means your existing content work carries over, but the deliverable format re-ranks your category in ways a chat does not. Treat 9 July 2026 as the start date for a surface you now need to watch, the same way you would after any model or picker change.
How to check your ChatGPT Work presence
Agent output is easy to overlook because you have to run a task to see it. Here is a short, practical way to find out whether ChatGPT Work names your brand:
- Give it a real brief. Ask Work to research and shortlist options in your category — the kind of task a buyer would set — and save the result with today’s date.
- See who makes the shortlist. Deliverables compress to a few named options; record which brands are included and which are left out.
- Open the sources. Ask which pages the research drew on — those are the places worth earning a mention.
- Compare agent against chat. Run the same question as a plain chat and as a Work task; differences show how the deliverable format re-ranks your space.
- Test connected context. Try the brief with and without relevant files or plugins connected, since context changes the shortlist.
- Repeat across regions and over time. The rollout is gradual, so one run is a snapshot — the value is in the trend as coverage widens.
How reconnAI tracks new surfaces
reconnAI monitors how the leading AI models — ChatGPT, Claude, Gemini, Perplexity, Copilot and Google AI Overview — answer questions across regions, and re-baselines that tracking whenever a platform changes how its answers are generated. An agent that researches and builds deliverables is exactly the kind of change that can move your standing on a surface you were not yet measuring. It lands the same fortnight as fresh model news elsewhere, from Claude Sonnet 5 to Perplexity’s Deep Research upgrade and Claude arriving in Microsoft 365 Copilot.
If you want to see how ChatGPT represents your brand now that Work can research and build on a user’s behalf, get in touch with our team or explore how AI visibility tracking works. And if your focus is the commercial side of AI answers, our overview of ChatGPT advertising covers where paid placement is heading.
ChatGPT Work: your questions answered
What is ChatGPT Work? It is an agent OpenAI launched on 9 July 2026 for longer, more involved tasks. It can research and analyse information, work across connected apps and files, and create finished documents, spreadsheets, presentations, reports and Sites, while you follow its progress and approve important actions.
When did ChatGPT Work launch? OpenAI began rolling it out on 9 July 2026, per the ChatGPT release notes, alongside a new desktop app and ChatGPT Sites.
Who can use ChatGPT Work? It is rolling out to paid plans except Free and Go, in all supported regions. Pro, Pro Lite, Enterprise and Edu get access first, with Plus and Business following over the coming days. Enterprise and Edu workspaces have a two-week preview during which it is off by default.
What is the Plugin Directory? It replaces the App Directory. Plugins package skills, apps and app templates for specific workflows, available from ChatGPT on web and desktop, including inside Work and Codex. Existing app connections are unaffected.
Does ChatGPT Work change how my brand appears in AI answers? It can. Work reads the live web to build deliverables and compresses a category into a few named options, so which pages it trusts and which brands it shortlists directly affect your visibility. Running your own briefs through Work is the way to tell.
About reconnAI
reconnAI tracks how the major AI models represent topics and sources across ChatGPT, Claude, Gemini, Perplexity, Copilot and Google AI Overview — across multiple regions. We monitor how those models answer and how they change over time, so you can stay ahead of shifts in the AI landscape.
Last updated: 15 July 2026.