Imagine paying for a premium AI subscription mainly to get work done with Codex. You try it, hit too much friction, and decide the upgrade is not worth it.
Then another feature walks in: Dots.
“Hang on. This one is actually useful.”
Congratulations. Your AI subscription has developed its own retention department.
Can Plus subscribers use Dots?
As of October 14, 2026, OpenAI does not list ChatGPT Plus as an eligible plan for Dots. The official rollout covers eligible Pro users, Business Premium, and Enterprise-family workspaces whose administrators enable the beta. No Plus launch date has been announced.
Here's the twist: Codex is available on Plus, although usage allowances and conditions differ between plans. You do not necessarily need Pro just to access Codex.
That makes the decision much stranger than a simple “pay or lose every tool” choice.
What is Dots, anyway?
It's pronounced “dots,” the plural of “dot.” Tiny name, surprisingly large job description.
A dot is an always-on AI agent. You give it an ongoing goal; it can use a dedicated cloud computer and connected apps to research, create, and make progress between conversations. OpenAI says it runs on GPT-6 Astra. You define what it may do autonomously, review the results, and step in when a decision needs your judgment.
In other words, this isn't merely a chatbot waiting politely for the next message. It's closer to a project helper that can keep a thread of work alive. The dot has somehow become a line of completed tasks.
Why can Codex disappoint while Dots impresses?
Product quality isn't a universal verdict. Some developers get excellent results from Codex; others may spend too much time correcting output, repeating instructions, or checking unfinished work. If that is your experience, questioning the cost is reasonable.
But the products offer different kinds of value. Codex often answers, “Can this coding task get done?” Dots raises another question: “Can this project keep moving while I'm doing something else?”
A good one-off assistant and a helpful ongoing agent are not the same proposition.
So the AI that gave you a reason to cancel may not be the AI making you hesitate. One created the exit interview; the other volunteered to conduct the stay interview.
How do you decide whether to keep Pro?
Measure completed work rather than excitement. Over a normal week, track what Dots actually delivered, how much supervision it required, and how much time it saved. A working result is worth more than an impressive demo.
If you mostly admire the potential, that's different from getting enough recurring value to justify the subscription. And before changing plans, check the current terms and your active tasks instead of assuming every feature or workflow will carry over unchanged.
The punchline
You opened the cancellation page because one AI underdelivered. Another AI appeared with a better pitch: “Let me handle the ongoing stuff.”
Brilliant retention strategy. Almost suspiciously brilliant.
Still, the subscription belongs to you. Even an always-on agent doesn't get to become the chief financial officer of your wallet.
