Most people’s ChatGPT account looks like a junk drawer after a few weeks of real use — dozens of stray chats, half-finished projects, context that never carries over. You finish a brainstorm in one thread, start a follow-up in another, and six weeks later you can’t find either one.
OpenAI’s new Dots feature is designed to fix that structural mess at the root.
What a Dot Actually Is
A Dot sits above your entire ChatGPT workspace — above individual chats, above projects, above all of it. Think of it as a persistent, always-on layer that can see and command everything underneath it.
Instead of hunting through threads to remember what you were doing, you open your Dot and ask. It knows. You can type to it or start a voice call directly, and it pulls context from across your account to answer.
It’s less like a chat and more like a chief of staff who has read every memo you’ve ever written.
The Problem It Solves
To understand why Dots matter, it helps to see how most people evolve their ChatGPT habits:
- Raw chats. You start a new thread every time you need something. Context resets constantly.
- Projects. You discover that grouping chats under a project lets them share instructions and background info. Better — but you end up juggling multiple projects, each with its own pile of threads.
- Dots. One place to interface with all of it.
The jump from step 2 to step 3 is significant. Say you’re running a freelance design business. You might have a project for client work, another for invoicing and admin, and a third for your own skill development. Today, switching between those means switching projects, re-orienting, re-reading. With a Dot, you just ask: “What’s the most urgent thing across my client projects right now?” and it tells you.
How to Use It
Your Dot appears in your ChatGPT sidebar, separate from your chat list. You can interact with it two ways:
- Text input — ask questions, give commands, request summaries
- Voice call — useful when you’re moving around or want to think out loud
Because it has visibility across your account, the kinds of questions that actually make sense to ask it are different from what you’d ask a single chat:
- “What have I been focused on this week?”
- “Which of my projects still needs a final review draft?”
- “Catch me up on the research I saved about competitor pricing.”
Those questions used to require you to dig. Now they’re just prompts.
What It Doesn’t Replace
A Dot isn’t a replacement for good project structure — it’s a reason to build that structure in the first place. If your chats are disorganized and your projects have no system instructions or shared context, the Dot doesn’t have much to work with.
Think of it this way: a Dot amplifies whatever organization you already have. A tidy workspace becomes genuinely navigable. A chaotic one stays chaotic, just with an extra interface layer on top.
So before leaning on Dots, it’s worth spending twenty minutes:
- Grouping related chats into named projects
- Adding a short context note to each project (who it’s for, what it covers, current status)
- Archiving anything stale
That upfront tidying pays off immediately, Dots or not.
Who Benefits Most Right Now
Dots are especially useful if you’re using ChatGPT across multiple domains of your life — work, side projects, personal research. The more contexts you’re managing, the more painful the constant context-switching is, and the more a single command layer helps.
If you mostly use ChatGPT for one-off tasks — rewriting an email here, generating a recipe there — Dots probably won’t change much for you yet. But if AI has become a genuine part of how you work every day, having one coherent interface to all of it is a real shift.
The underlying idea is simple: your AI workspace should work the way your brain does — holistically, not in isolated silos. Dots are OpenAI’s first real push in that direction.