ChatGPT dots: an ongoing AI helper for research, drafting and follow-up.

ChatGPT dots Explained: Features, Everyday Uses and Your First Safe Workflow

Read ChatGPT dots Explained: Features, Everyday Uses and Your First Safe Workflow on LinkedIn

You run a course, an agency or a small team. The work keeps changing after you close ChatGPT: a launch date moves, a proposal changes, another question needs an answer.

ChatGPT dots is designed for that ongoing work. You give your dot a responsibility, relevant sources and boundaries. It can research, prepare outputs and follow up between conversations. Here is a visual guide to deciding what to delegate first.

Product facts checked October 1, 2026. The examples below illustrate documented capabilities; they are not tested client results or adoption statistics.

1. What is a dot?

Define the goal, choose sources, work in the background, review the result and confirm the next step.
A proposed responsibility-to-review workflow, not proof that a schedule is already configured.

A dot is an always-on agent with its own cloud computer and browser, powered by GPT-6 Astra. It can retain relevant context and coordinate background tasks. Think of an ongoing assignment: “Keep this launch brief consistent as our approved plans change.”

You still define success and review important outputs. Memory is useful context, not a perfect record of everything. OpenAI overview · Tasks and memory

2. Does my laptop need to stay on?

Cloud work can continue with your laptop off. Personal-computer work requires it online with ChatGPT open.
Cloud execution and optional personal-computer execution have different requirements.

For cloud work, no. For work on your connected personal computer, yes. The personal computer must be online with the ChatGPT app open. Connecting it is optional, and only one personal computer can be connected at a time.

The cloud browser also has separate sign-ins. Your local logged-in tabs do not automatically become accessible. Some websites may block cloud browsers. Computers and apps

3. Where can I talk to it?

Messaging controls where you talk; apps determine accessible data; computer connections determine where actions run.
A messaging connection does not automatically grant app or computer access.

ChatGPT, Slack and Microsoft Teams are documented messaging channels. Voice is available too. But connecting a messaging channel does not connect your CRM, inbox or computer. Each needs its own supported connection and permissions.

Texting is coming soon; dot-initiated calls are planned after launch. Do not build a customer promise around either yet. Channels

4. What could I use it for?

Suggested uses include proposal upkeep, content drafts and launch checks.
Suggested examples based on documentation, not measured customer outcomes.

OpenAI's examples include keeping sales proposals current, turning interviews into content drafts, checking launch-material consistency and turning software feedback into proposed fixes.

For a course business, a sensible adaptation is a launch-review assistant: compare your approved course brief with supplied landing-page and email copy, then flag conflicting dates, access promises or missing instructions. It drafts corrections; you approve them.

These are starting points, not proof of a native integration with every course platform. Documented examples

5. Give it a brief it can actually follow

A useful brief defines a goal, approved sources, allowed and forbidden actions, and timing plus review trigger.
Confirm sources, boundaries and saved recurrence before relying on ongoing work.

Try this low-risk example:

“For the next seven days, review the launch files I supply at 9 AM Asia/Dhaka. Compare them with the approved course brief. Report changed dates, contradictory access promises and missing next steps in ChatGPT. Link each issue to its source. Draft corrections only. Do not edit live pages, contact students or change CRM records. Confirm the schedule, sources and limits before starting.”

A request alone is not proof that recurring work was configured. Check its confirmation and Scheduled. Getting started

6. Keep the review and stop controls separate

Pause controls the main task; inspect delegated tasks in Activity and future runs in Scheduled. Stopping is not undo.
Main tasks, delegated work and recurring schedules need separate review.

Use Activity to inspect delegated work and results. Review recurring work in Scheduled. Pausing the main task does not automatically stop delegated tasks or cancel future runs. Completed actions are not undone by stopping.

Custom permission rules can define review boundaries, but they do not grant app access or override built-in safeguards. “Draft a reply” should never mean “send it.” Controls

7. Can I use it today?

Start with one responsibility, read-only sources and one reviewed report. Expand only if useful.
A suggested low-risk pilot, not a performance guarantee.

Availability is rolling out gradually. OpenAI lists Pro 100, Pro 200 and Pro 500 for adults outside the EEA, UK and Switzerland; Business Premium worldwide; and Enterprise worldwide with admin enablement. Eligibility does not guarantee it appears immediately. Mobile web is unsupported; mobile-app availability requires a supporting update.

Dot conversations and delegated Work/Codex tasks have different usage treatment. Do not assume unlimited execution. Check your account and current availability.

My recommendation: start with one read-only responsibility and one useful report. Measure whether it catches a real issue and reduces your review effort before adding more access.

Which recurring task would you delegate first: launch checks, proposal updates or content preparation?

AI-assisted article and conceptual illustrations, checked against the official sources linked above. No hands-on workflow benchmark is claimed.

Back to blog