ChatGPT Space is a shared team hub, Pages is a collaborative document format and Team Tasks delegates recurring work to connected tools on a schedule or trigger, all announced by OpenAI on 29 September 2026. For a small agency, the combination is promising for status reports, briefs and recurring admin, but it is aimed at Business and Enterprise plans, and I have not run it in production. This review is based on OpenAI's announcement and press coverage as of 30 September 2026, so treat the workflow below as a plan to test, not a proven result.
Key Takeaways
- Per OpenAI's recap, Space is for Pro, Business and Enterprise (desktop and web first); Pages is for Pro, Business and Enterprise; Team Tasks and @ChatGPT in Slack and Teams are for Business and Enterprise.
- Team Tasks handles recurring work such as weekly updates and email responses using connected tools, on schedules or event triggers.
- OpenAI says teammates can add context via @ChatGPT in Slack or Teams without needing an individual ChatGPT licence.
- Engadget reports Space is rolling out first on desktop and web, then mobile, and Pages is coming in the weeks ahead, with slide export to PowerPoint and Google Slides.
- Small agencies should start with internal workflows (status reports, briefs, meeting follow-ups) before client-facing ones.
- Client confidentiality, permissions and review steps are the real work.
What each feature is
Based on OpenAI's DevDay 2026 recap:
- ChatGPT Space: a collaboration hub that brings people, AI and agents together on shared knowledge, with ChatGPT organising content per your instructions.
- Pages: documents for writing, research, charts, images and visualisations, with real-time team contribution and feedback.
- Team Tasks: delegation of recurring work through connected tools on schedules or event triggers, for Business and Enterprise.
- @ChatGPT in Slack and Microsoft Teams: mention it in channels, threads or DMs, for Business and Enterprise.
Engadget's report on Space adds that Space rolls out first on desktop and web, then mobile, Pages arrives in the coming weeks with collaborative slides exportable to PowerPoint and Google Slides, and that meeting notes are available on macOS for Pro and Business users. That article did not describe Team Tasks, so my detail on it comes from OpenAI's recap.
Who this actually fits
A three-to-fifteen person agency has three problems these tools might address: knowledge scattered across drives and chats, repetitive reporting, and slow handoffs. The fit is reasonable, with two caveats. First, Team Tasks needs a Business or Enterprise subscription, which is a cost step for very small teams. Second, agencies handling many client accounts must decide how to separate client data.
A sample agency setup
Below is the structure I would pilot. It is a design, not a case study.
One Space per client
Keep client knowledge separate: brand guidelines, past reports, keyword lists, approved tone, and a short "do not do" list. A Space per client reduces accidental cross-contamination of confidential material. I would name spaces consistently, for example "Client A – SEO".
Pages for living documents
Use Pages for briefs, content calendars and monthly report drafts that several people edit. The advantage is collaboration on AI-generated text and charts in one place. Export to slides is helpful for client presentations, once available.
Team Tasks for recurring admin
Candidate tasks, each with a human review before anything leaves the building:
| Recurring job | Trigger | Output | Review step |
|---|---|---|---|
| Weekly status update draft | Friday schedule | Draft in the client Space | Account manager edits and sends |
| Meeting follow-up summary | After a call | Action list in Pages | Owner confirms actions |
| New-lead inbox summary | Daily | Digest in Slack | Sales lead reads |
| Competitor content scan | Weekly | Short brief | Strategist checks sources |
| Reporting data pull | Monthly | First-draft report | Analyst verifies numbers |
@ChatGPT in Slack for quick asks
The recap says teammates can add context without needing their own licence. In an agency this could let a junior or a freelancer ask for a draft without a full seat, but check licensing terms so you are not breaching any contract.
The trade-offs to weigh
Cost and seats
Business and Enterprise plans cost more than an individual plan. Confirm current pricing on OpenAI's site, because I did not verify it. For a small team, calculate against the hours saved on recurring reports.
Confidentiality
Agencies sign NDAs. Before you put client data into any shared AI workspace, read the plan's data terms and your contracts. Some clients ban third-party AI processing outright. Ask first.
Permissions and access
Shared workspaces are only as safe as their access controls. Decide who can see which Space, and remove contractors' access when projects end.
Review discipline
Team Tasks that draft weekly updates can generate plausible mistakes: the wrong week's numbers, a misread trend, an invented explanation for a traffic dip. The rule I would enforce is that no AI-drafted client communication goes out without a named human's edit.
Dependence
If your process lives inside one vendor's workspace, exporting it later can be hard. Keep master copies of client documents in your own storage.
A four-week pilot
- Week 1: Create one internal Space and one client Space with non-sensitive material. Set access rules and write a two-paragraph AI use policy.
- Week 2: Turn on one Team Task, the Friday status draft. Compare its output to your manual version and log errors.
- Week 3: Add Pages for one monthly report draft. Time the process against your baseline.
- Week 4: Review the log. Keep tasks where drafting time dropped and errors were rare. Cancel the rest.
Measure hours saved per week, errors caught in review and team satisfaction. If you cannot show saved hours, do not upgrade seats.
What is still unknown
- Exact usage limits and pricing per seat for Team Tasks.
- How well Team Tasks handles connectors that require complex permissions.
- How reliably the scheduled runs complete, since nobody has published failure-rate data.
- How Space handles large client knowledge bases.
- Whether guest or client access is possible; I did not find this in the sources I read.
How it compares to what you already use
Many agencies already run Slack, Notion, Google Docs and project tools. The value of Space and Pages is not that they are better documents; it is that the AI has your context in one place. If your team already uses another AI-integrated workspace, test whether switching gains enough to justify the migration cost. Sometimes it will not.
Sources checked
- OpenAI, DevDay 2026 Recap (primary; features and plan availability).
- Engadget, ChatGPT Space is a shared hub for work projects and personal AI agents (rollout order, Pages, meeting notes).
- OpenAI Help Center, Teams in ChatGPT (appeared in search results; I did not read it in full, so confirm details there).
FAQ
What is ChatGPT Space?
Space is a shared hub where teammates, ChatGPT and agents work with shared files and knowledge. OpenAI lists it for Pro, Business and Enterprise plans, rolling out first on desktop and web.
What are Team Tasks?
Team Tasks delegates recurring work, such as weekly updates or email responses, to ChatGPT using connected tools. It runs on schedules or event triggers and is listed for Business and Enterprise plans.
What is Pages in ChatGPT?
Pages are collaborative documents where you can generate text, charts, images and visualisations and share them for team feedback. Engadget reports slide export to PowerPoint and Google Slides is coming.
Can teammates use @ChatGPT without a licence?
OpenAI says teammates can add context in Slack or Teams without needing an individual ChatGPT licence. Check the plan terms for what is allowed, especially for contractors and clients.
Is this suitable for a two-person agency?
Possibly, if recurring reports take real time each week. Business plan costs and the setup effort may not pay off for very small teams. Run a short pilot and measure hours saved.
Is it safe for client data?
It depends on the plan's data terms and your client contracts. Read both before uploading anything confidential, and apply strict access rules to each Space.
Which Team Tasks should I start with?
Internal, low-risk ones: a weekly status draft, meeting follow-ups and inbox digests, each with a human review step. Avoid automatic client-facing sends.
Does it replace Slack, Notion or Google Docs?
Not automatically. It adds an AI-native layer with shared context. Test whether it beats your current setup on time saved before migrating anything.
Are there limits I should know about?
I could not find published usage limits or per-seat pricing for Team Tasks in the sources I checked. Confirm them on OpenAI's site before committing.
Work with me
I have 4+ years of marketing experience in SEO, content and organic social growth, including edtech work with Masai School. If you run a small agency and want an honest look at where AI workflows will actually save time, see my work and contact me through the form at younusfardeen.in.