OpenAI dots launch in Australia, with a first dot included

OpenAI is adding an ongoing assistant to eligible Australian ChatGPT subscriptions, with the first dot included at no additional charge. For an owner already paying for Pro or Business Premium, that makes a small trial worth considering. The harder decision is how much of the office to let it into.

In its 29 September announcement, OpenAI describes dots as agents powered by GPT-6 Astra, with their own cloud computer and access to connected apps. The company says they can keep working towards a goal between conversations and bring back results for review. Those are launch claims, rather than evidence of how much checking a small business will need to do.

The useful question is whether handing over one repeated job saves the office manager time once review and corrections are counted. An assistant that keeps working needs a brief that says where to stop.

Who can try a dot in Australia

OpenAI’s setup guide says Pro access excludes the European Economic Area, Switzerland and the UK, while Business Premium access covers supported ChatGPT regions. Australia is on the supported-country list, so on our reading it is within both rollout areas. Individual accounts may still be waiting: OpenAI says access can take several days.

Business Premium is the eligible business tier named in the launch. Enterprise users, including Edu and Healthcare, can try a beta if their workspace administrator enables it; it starts switched off.

For existing eligible Pro and Business Premium subscribers, the additional charge for the first dot is zero. We have not verified an Australian-dollar base subscription price or checkout total. A business considering a new subscription should check its local checkout before deciding; this launch alone is not a reason to upgrade on an assumed price.

Included access also has limits. OpenAI says plans have an allowance for deeper work, temporarily extended during the first launch month. Dot conversations do not consume ordinary ChatGPT usage limits, but tasks the dot starts or manages in Work or Codex consume those task limits. The launch offer should not become the business’s assumption about its ongoing capacity.

Check the apps already connected

Create a dot through the ChatGPT desktop app or desktop web. Mobile cannot currently create one, and mobile web is unsupported. Its cloud computer is separate from yours; optional access to your own computer starts disabled, according to the setup guide.

Connected apps deserve attention before business work begins. OpenAI’s privacy and safety FAQ says plugin permissions are shared across dots, ChatGPT, Work and Codex. A dot can use existing connections within the access already granted. Review them in ChatGPT’s Plugins tab rather than assuming a new assistant starts with no access.

OpenAI says Business, Enterprise and Edu data is not used for model training by default. On personal plans, the model-improvement setting controls eligible use. Disconnecting an app stops new access but does not erase information the dot has already retained. That gives the owner a reason to settle the data rules before connecting another business account.

Our small-team AI policy guide can help record which accounts and information are approved. Keep that decision with the person responsible for the business, even if a staff member is keen to try the feature.

Give it one job and an end date

Our recommendation is a one-week trial using public or otherwise approved information. For example, ask the dot to check one nominated supplier’s public announcement page each morning, flag changes relevant to your business and link to the original source. This is an illustrative brief, not a workflow we have tested.

Specify the page, what counts as relevant, who reviews the findings and when the checks end. Keep the trial to reporting in your conversation, with no authority to email customers, change records or buy anything. Record the time spent checking against the time the same job normally takes. Our guide to choosing a first AI task you can check explains that approach.

OpenAI says Custom Rules can add supported action boundaries, but cannot override core safeguards. Its FAQ also says approving one message does not grant ongoing permission to contact people. Review consequential work: the company warns that dots can still make mistakes.

The first dot’s inclusion removes an extra purchase decision for eligible subscribers. A useful trial still needs an owner, a narrow remit and a result someone can verify. More AI at work coverage follows the product changes; expanding this trial should depend on what the business actually learns.

Checklist

• Check eligibility in your ChatGPT account. Confirm Pro or Business Premium access and allow for the gradual rollout before planning the trial.
• Review connected apps in ChatGPT’s Plugins tab. Have the account owner check existing permissions before starting business work with the dot.
• Write the one-week trial brief. Name the public source, relevant changes, reviewer, reporting destination, approval boundaries and end date.
• Compare checking time with the original task. Have the named reviewer record corrections and review time before the owner decides whether to expand the work.

Artwork update, 1 October 2026: We connected this story’s illustration after a missing image reference left the first published version showing section artwork.