---
title: 'ChatGPT Dots Just Got a HUGE Upgrade! 🤯'
source: 'https://youtube.com/watch?v=Ee3y7zXTQfw'
video_id: 'Ee3y7zXTQfw'
date: 2026-11-17
duration_sec: 481
channel: 'Julian Goldie SEO'
---

# ChatGPT Dots Just Got a HUGE Upgrade! 🤯

> Source: [ChatGPT Dots Just Got a HUGE Upgrade! 🤯](https://youtube.com/watch?v=Ee3y7zXTQfw)

## Summary

This video explains OpenAI's new 'Dots' feature in ChatGPT, an always-on agent that can handle ongoing tasks, and details a recent update that adds custom instructions and rules. It covers how to set up these features, common mistakes to avoid, and who currently has access.

### Key Points

- **Introduction to Dots** [00:03] — Dots are OpenAI's always-on agents inside ChatGPT, launched on September 29, 2026, at DevDay. They run on their own cloud computers with a browser, can work while your laptop is off, and run on GPT-6 Astro.
- **November 12th Update** [01:28] — The official ChatGPT account announced three improvements: faster replies, less jargon, and better follow-through. However, no specific numbers or tests were provided.
- **Custom Instructions for Dots** [01:43] — Dots can now follow custom instructions, allowing users to set ongoing behavioral preferences. Instructions are separate from regular ChatGPT custom instructions and must be set specifically for the dot.
- **Three Ways to Give Instructions** [02:38] — Users can give instructions via chat (e.g., 'Always talk in lowercase'), mobile (sidebar > customize > behavior), or web/desktop (dot's menu).
- **Best Practices for Instructions** [03:37] — OpenAI recommends giving each instruction a clear scope: who, what, and when. This helps the dot understand when to apply the instruction.
- **Custom Rules vs. Instructions** [04:03] — Custom rules control when a dot can take actions (e.g., sending emails), while instructions are for preferences. Rules are set in settings > personalization > custom rules.
- **Common Mistake: Draft vs. Send** [04:30] — Asking a dot to write a draft does not give it permission to send it. Users must explicitly set up a draft-first workflow if desired.
- **Availability and Limits** [05:13] — Dots are rolling out to ChatGPT Plus users 18+ outside the EEA, UK, and Switzerland, plus eligible business workspaces. Chatting with dots doesn't count toward usage limits on Pro and Business Premium, but tasks in Work or Codex do.
- **Additional Controls** [06:05] — Adding a dot to Slack doesn't auto-watch channels; you must specify. Pausing only stops the main task, not handed-off or scheduled tasks. Stopping doesn't undo completed actions.

### Conclusion

Dots represent a significant step toward autonomous AI agents, but users must carefully configure instructions and rules to avoid mistakes. The update makes them more customizable, yet safety and oversight remain essential.

## Transcript

Ever told your AI the same thing 10 times? Yeah, me too. Well, ChatGPT dots just got a fix for that, and it changes how you work with them. Most people haven't even noticed yet. Stick around because later in this video, I'll show
you the one setting that's really easy to miss, and the one mistake that could catch you out if you skip it. So, first, what even is a dot? A dot is OpenAI's always-on agent inside ChatGPT. It launched on September 29th, 2026 at
OpenAI's DevDay. Here's the simple version. Normal ChatGPT waits for you to ask something. A dot doesn't just wait. It takes on ongoing work, keeps track of it, and checks back with you when a decision needs you. Each dot runs on its
own computer in the cloud with its own browser, so it can keep working even when your laptop is off. It runs on OpenAI's GPT-6 Astro model. It can also run background agents on a few things at once while you keep chatting with it.
And you can connect your own computer through your dots profile, so it can use your local files and apps. And you can reach it in a bunch of places. You can message it or call it in ChatGPT on desktop, web, and mobile. You can
message it in Slack, and you can message it in Microsoft Teams. Switch channels, and it's still the same dot with the same memory. Okay, so that's dots. Now, here's what just changed. On November 12th, the official ChatGPT account on X
posted an update. It said your dot now has three things, faster replies, less jargon, and better follow-through. Now, one honest note, OpenAI didn't share any numbers or tests for those three. So, that's what OpenAI says. Your own
experience might feel different. But the biggest part of this update is the next bit. Dots can now follow your custom instructions. That means you can tell your dot how to behave, and it keeps that as an ongoing instruction, not just
for one message going forward. Quick pause before I show you the part. Inside the AI profit boardroom we've got a 30-day AI agents roadmap. It walks you so you're not trying to figure everything out in one go. There are also
four coaching calls every week. So if you get stuck setting up something like a dot you can bring your exact question to a call and work through it live. That way you're not guessing. You're learning it properly one step at a time next to
other people learning the same tools. Link in the description or go to ai profit boardroom.com. All right, back to dots. So how do you actually give your dot instructions? There are three ways. Way one, just tell it in the chat. In
your dots thread say what you want. OpenAI gave two examples in their post. Always talk in lowercase and show me a draft before sending any emails. When you finish a task link to the result and tell me what needs my
attention and if you want to change it later just ask your dot to update the instruction. Way two, on mobile open the sidebar, tap customize then tap behavior. If you see tabs for ChatGPT and your dot pick your dot's name, type
in your preferences and hit save. If you don't see behavior update your app or use your dot's chat on desktop or web instead. Way three, on web and desktop. dot's menu. Now here's the setting that's easy to miss. Remember I
instructions are separate from your normal ChatGPT custom instructions. So if you already set up custom instructions in ChatGPT don't assume your dot got them. It didn't. You need to set them for your dot. So what should
you put in there? OpenAI says to use instructions for things like writing style, how often it checks in and when it should ask before acting. And here's a top tip straight from OpenAI's help docs. Give each instruction a clear
scope, who, what and when. Spell out who it applies to, what it covers and when it kicks in. Next, custom rules. These are different from instructions and this is where people get mixed up. Instructions are for preferences, your
style, how you like updates. Custom rules control when your dot can take an action. On desktop and web, go to settings, then personalization, then custom rules under permissions. On iOS and Android, open the sidebar, then
customize, then custom rules. Tap add, describe the action, then pick one of four options. Now here's the mistake I promised to warn you about. Asking your dot to write drafts does not give it permission to send them. Those are two
different things. So if you want a draft first workflow, set it up clearly. Also, neither instructions nor custom rules override OpenAI's built-in safety requirements. Rules don't give your dot access to apps it doesn't already have.
And OpenAI says rules are instructions the dot tries to follow, but it can still make mistakes. So check anything important. On top of that, there's an something that could affect your accounts or share information, it gets
checked against your instructions, permissions, rules, and safety requirements. Then one of three things happens. It goes ahead, it needs your approval, or you have to do the step yourself, like changing a password. Now
who can actually get a dot right now? Dots are rolling out gradually. Right now, that's ChatGPT Plus users who are 18 or older outside the European Economic Area, the UK, and Switzerland, plus eligible business premium
workspaces. Enterprise, EDU, and healthcare workspaces can get them too. And from November 18th, 2026, access turns on by default in those workspaces unless the workspace opts out. You can create your dot in the ChatGPT mobile
app, the desktop app, or a desktop browser. Mobile web isn't supported. At desktop, so mobile setup is a nice addition. Good news on limits too. On
Pro and Business Premium, chatting with your dot doesn't count toward your ChatGPT usage limits. But if your dot starts tasks in ChatGPT Work or Codex, those count toward those limits as normal. A couple more controls worth
knowing. If you add your dot to a Slack channel, that doesn't start it watching the channel. You need to tell it which channels and events to watch. If you want an update at a set time, include your time zone and a duration or end
date. Then ask your dot to confirm the schedule. And if you're a pro user in the US, texting your dot is in beta too. On desktop, open your dot's profile and tap activity. You'll see task progress, files, and results, including background
work. And if you want to stop things, know this. Pausing only stops the current main task. It doesn't stop tasks your dot handed off, and it doesn't cancel scheduled runs. To end a recurring task, go to scheduled and turn
it off or delete it. Stopping a task also doesn't undo things it already did. So set your instructions and rules before you hand over important work, not after. Here's the thing. There's a window right now where most business
owners haven't figured this out yet, and that window won't stay open forever. dots now means you'll know how to use them well. Inside the AI Profit Boardroom, there are daily tutorials that walk you through new AI tools as
calls every week where you can bring your questions and get help live. There's a prompt library, so you've got ready-made prompts to learn from. And there's the member map so you can find and connect with other members. Link in
aiprofitboardroom.com. If you want the full process, the SOPs, and over 100 AI use cases like this one, join the AI Success Lab. It's our free community. Links in the comments and description. You'll get all the video
notes from this episode plus access to 95,000 members. The AI Success Lab is completely free to join. The video notes give you a written breakdown of this episode so you can go back over every step without rewatching. And you'll be
learning alongside a huge community of people all getting better with AI. So people all getting better with AI. So come join us and I'll see you inside.
