ChatGPT Dots: Pointless or Priceless? Everything You Need to Know
AI for Real People
OpenAI's Dots are an always-on assistant that lives inside ChatGPT, and the price tag has three digits. After about two weeks with mine, here's what actually works, what doesn't, and who should skip it.
What happens to your work when you disappear for a while? That's the question behind Dots, OpenAI's new always-on assistant. It's built to keep your projects moving, use the apps you connect, and report back. I've been living with mine for about two weeks, and I have opinions.
Here's the problem I keep running into: the pitch sounds like a whole team of little assistants, and the product is one assistant with some helpers underneath.
Whether that's worth $100 a month or more depends on how you work.
What is a Dot, exactly?
In OpenAI's words, a Dot is more or less an always-on browser that can use the apps you connect. You give it a goal, you tell it what it's allowed to do on its own, and you tell it where you need to be pinged. That's the product, and it's a good product.
It also comes with some pros and some annoying cons. Enough of them to stop me from moving all my workflows over to my Dot.
What actually works?
1. The voice call
This is the closest thing I've found to an actual coworker. You start it from your ChatGPT app on your computer or mobile. Open your Dot's conversation, press the phone button, and you start talking.
2. One Dot, with helpers underneath
You get one Dot, and it can run helpers underneath it. Those helpers aren't Dots. You can't message them directly and you can't buy more of them. Hold that thought, because it comes back below.
3. Phone or desktop, it's the same Dot
Your Dot in the ChatGPT mobile app is the same one you get on desktop, with the same preferences and decisions. Those notes are separate from ChatGPT's saved memory. And like any other AI memory out there, it's not a full transcript of everything you said and agreed to. It's summaries.
4. Your apps, as usual
You can connect your calendar, email, Slack, Notion, Obsidian and more from the plug-in list. This is where "what needs my attention today?" actually works. Your Dot can read across all of them and send a Monday check-in as a recurring event. That's what I have: one morning we jump on a quick call and I get an update on everything that needs me.
When you're not talking to it, it also does what OpenAI calls proactive research. It reads your connected apps and looks for ways to help. It's read-only, so it can't send a message or change anything while you're asleep. It can only come back with suggestions. In theory, the more you use it, the more you get back. In practice, it's too early to tell.
On desktop you also get custom rules: take action without asking, take action if pre-approved, ask before taking action, or hand off to you. Set those before you connect your email or anything sensitive. Keep that in mind.
What doesn't work yet?
1. You only get one Dot
I'm not going to lie, this was a little disappointing. OpenAI's launch post says you can start with your primary Dot, and that you'll be able to add more in the future. More Dots are promised. They're just not for sale right now, so no, you can't buy a pack of three. That's quite a limitation, especially if you're used to working with Grok Bots, which is a different situation.
2. It takes Pro money, and Pro Dots don't ship everywhere
Pro starts at $100 a month. Then there are higher tiers at $200, and at $500 and above. Pro is monthly only, no annual discount, in case you were wondering. On the business side, the Premium seat is $100 a month on annual billing or $125 a month month-to-month, and a business workspace starts at 2 seats minimum.
Your Dot comes with a three-digit price tag, if it's available in your region. If you're in the European Economic Area, the UK or Switzerland, Dots aren't available on personal Pro plans at all. Business Premium gets Dots in all supported regions, so if you're one of those places and you want a Dot, the only way today is a Business Premium seat, which means a 2 seat workspace minimum. Do that math before you get really excited, because as it stands, I don't think Dots is really worth it.
3. You don't really get a team
This is the one I keep hearing people complain about, and I complained about it a little at the beginning too. The initial public perception was that your primary Dot could summon a team of other Dots, and that you could have multiple primary Dots with dedicated roles. When I asked my Dot about it, it gave me the same impression. That's not what happened in execution.
People want one dedicated agent per job: one for research, one for client follow-ups, one for marketing, one for packaging the week. You can't do that today. You get one Dot that can run multiple helpers underneath it through ChatGPT Work or Codex, but a helper isn't the same as having your own owner. It has no memory of its own, no rules of its own and no channel of its own, and you can't chat with it directly. If you fire a question into Slack at 11 at night, the same Dot answers, not the little helpers underneath.
That's different from Grok Bots, where you can have dedicated bots that work together in a group, share knowledge and collaborate. More on that next week. Subscribe so you get notified when it's out.
4. Usage is confusing, and "your Dot is on a break" has no meter
OpenAI's launch post says conversations with your Dot don't count toward your ChatGPT usage limits. Good, right? You can talk to your Dot all day. The same paragraph says that when you ask your Dot to manage tasks in Codex or ChatGPT Work, those tasks do count toward your usage limits as usual. So the helpers underneath come out of your normal limits.
Then there's a third sentence people skip: your plan includes an allowance for deeper work with extended limits for the first month after launch. First month only. That's not forever. It means October is really generous, and OpenAI hasn't told us what it looks like after that, so take advantage of it.
Which brings me back to the "on a break" situation. A Pro user on OpenAI's own community forum, paying for the $200 plan with 47% of their usage still showing, asked their Dot to build a job tracker in Google Sheets overnight. They woke up to a message that their Dot was on a break: congrats, you're among the very top users and have hit our abuse prevention limit. Three hours later, their Dot was still on a break.
So Dot chat might not touch your normal ChatGPT meter, and you can still get parked with no bar on the screen telling you how close you were or when the break will end. Two separate systems, one of them a little bit invisible.
Is a Dot worth it? My yes and not-yet list
Yes, for:
Brainstorming on a voice call. It does this better than anything else I've tried. You think out loud, change your mind, interrupt, ask questions, hang up and come back. It ties the thinking and the work into something written and structured that you can move on from. I also enjoyed just going for a walk while brainstorming with my Dot, and that was incredibly productive.
The morning attention check. I have multiple calendars and inboxes across multiple time zones. A read-only update on what actually needs my full attention today saves me quite a bit of time, and it can't send anything without my consent.
Small jobs under one Dot that report back without spinning up complicated helpers: one research task, one draft of an email or document, one "go check this and let me know."
Not yet, for:
One dedicated Dot per job. If your workflow needs separate owners, Dots just aren't for you. (We should probably stop calling them Dots and start calling it a Dot, because that's what it is at the end of the day.)
Anything with a deadline that runs overnight and unattended, until the break screen gets a meter. It's not that it won't work. It's that you won't know when it's going to stop, and no one seems to know when exactly those breaks kick in.
What I actually built with it
I used mine to brainstorm an AI-agnostic skill library. The result was quite decent, though I still needed to make some improvements and refinements. The solution it recommended was along the lines of what I was thinking in the first place: a library of skills written simply in Markdown, plus an adapter for each platform I use. No skill duplication, and if there's a new AI tool tomorrow, I can bring my skills and stack over quickly. I have one adapter for Grok, one for Claude and one for ChatGPT.
It reminds me of the early days as a software engineer, when you had to set up everything from scratch for every project and boilerplates became a big thing. That's where we are now, with your harness, your skills, your operating principles and your standards. You have to keep moving them between tools, and I like to test them all, so I need something AI-agnostic. Otherwise you're locked into one AI tool, which is dangerous this early.
For You:
If you're already on Pro or Business Premium in a region that has Dots, try one thing this week: connect your calendar and email, set your custom rules to "ask before taking actions" first, and take your Dot for a walk. Press the phone button and brainstorm one real problem out loud. Then tell me what annoyed you.
If you're not on one of those plans, don't buy it yet. Send this to the friend who's about to drop $100 a month or more on a "coworker fantasy," a.k.a. Dots, before they know what one Dot actually is. Save them some money in this economy.
The full walkthrough is on YouTube: The Casual Leader, "ChatGPT Dots: Pointless or Priceless? Everything You Need to Know."
If it's useful, like, subscribe and hit that notification bell so you never miss an episode.
We have tons of interesting subjects coming.
If you're trying a Dot this week, let me know at contact@thecasualleader.com what you're using it for and what annoys you the most. I read and reply to every single one.
And join our free community newsletter over at thecasualleader.com for more tech and AI updates, business challenges, behind the scenes and more.
Start casual. Keep going. Keep growing.
