Does a ChatGPT dot replace a personal CRM?

You have heard that ChatGPT now has dots: always-on agents that remember context and work across the apps you connect. If you keep a personal CRM, a tool for keeping track of people and what was said, you may be wondering whether a dot replaces it.

Going by OpenAI’s own pages as we read them on 30 September 2026, the two do different jobs and overlap in the middle. A dot is a general agent you hand work to. A personal CRM is where you look up a person and read what was actually said.

The clearest difference is in OpenAI’s FAQ for dots: “You currently cannot view, delete or directly modify individual dot memories, including specific details that enter the dot’s context from plugins.” [1] As documented today, you cannot open what a dot has learned and read it item by item.

What a dot is

OpenAI’s help page describes a dot as “an always-on agent in ChatGPT that can take on ongoing responsibility and keep making progress between conversations.” It has its own cloud computer and, the same page says, “works across the apps you choose to connect, and remembers context to help with ongoing work.” [2]

You can ask it to set a reminder or run a recurring check, such as reviewing your calendar each morning. [2] For now there is one per person. The announcement says “you can start with your primary dot, give it a name, and make it your own.” [3]

What it remembers, and where that comes from

OpenAI’s help page says: “Your dot receives memories from ChatGPT and can create its own memories, including from connected apps.” [2] The sharing with ChatGPT runs both ways: the FAQ says conversations with your dot “can contribute to memory in ChatGPT.” [1]

A dot also forms memories without being asked. Of the information in connected apps, the help page says a dot can “review that information proactively and form memories from it, even when you haven’t asked a specific question about it.” [2] The FAQ says it can keep context from your conversations and plugins (OpenAI’s other word for connected apps) “for as long as you keep your dot”. [1]

This is the overlap with a personal CRM. Both build on the apps you connect, and both carry context forward.

What you can and cannot inspect

Beyond that limit on individual memories, which OpenAI describes with the word “currently” [1], this is how the controls are documented today:

• The Memory screen. On mobile, the help page says, “Customize → Memory shows what ChatGPT remembers. This screen uses shared ChatGPT memory settings.” It adds: “To delete your dot’s own saved memories, you need to delete your dot.” [2]
• Disconnecting an app. “Disconnecting an app does not delete information your dot has already obtained from it.” [2] It stops new access through that connection. [1]
• Turning off Memory in ChatGPT. This stops the sharing between ChatGPT and the dot. “It does not delete information that your dot has already received.” [1]
• Reset. “Reset deletes your dot, including its conversations, saved memories, and scheduled tasks.” [2] The FAQ lists what deleting a dot leaves in place: files, Codex threads and ChatGPT conversations the dot created, and the memories from your other ChatGPT conversations. [1]

Ordinary ChatGPT memory is documented differently. OpenAI’s memory help page says: “You can tell ChatGPT that remembered information is incorrect or outdated. You can also edit the memory summary or a saved memory when those controls are available.” [4] Our guide to correcting an AI memory walks through those controls.

For a dot, the documented way to fix a mistake is to steer the work. Activity View in the desktop app lets you “add context, correct a misunderstanding, change direction, or ask a dot to stop.” [5] The pages we read do not say what a correction like that does to a memory the dot has already stored.

Other people

A personal CRM is mostly about other people. OpenAI’s pages describe no separate record of the people you deal with, and no controls just for what a dot learns about them. Those details come under the general rule quoted above, which covers “specific details that enter the dot’s context from plugins.” [1]

Email and Slack are largely made of other people’s messages, so memories a dot forms from them are likely to include details about those people. That step is our reasoning; OpenAI does not spell it out. If a dot had picked up something wrong about a colleague, the documentation describes no way to find that one item and change it.

We cover the same question for ordinary assistants in what AI assistants remember about the people you mention.

Reading and acting

When you are not working with it, a dot does what OpenAI calls “proactive research”, using your connected apps with tools “restricted to be read-only, which means that they can’t send messages, change app content, or control your browser or computer.” [3] The research still saves private notes for the dot. [5] Work you have asked for or scheduled is a separate case: the FAQ says the usual action rules apply “to tasks you’ve already authorized that continue in the background.” [1] OpenAI says dots start with built-in rules “for when to act independently and when to ask for approval”, and you can add Custom Rules of your own. [3] Our article on what an AI agent may read and what it may do has the detail.

Who can get one

As documented on 30 September 2026:

• Pro. Rolling out “to Pro users in markets excluding the European Economic Area, Switzerland, and the UK.” [2]
• Business Premium. Available “across all supported ChatGPT regions.” [2]
• Enterprise, including Edu and Healthcare. A beta that a workspace admin has to enable; it is “initially turned off by default.” [2]
• Age. “Dots are not yet available to users under 18.” [1]
• Setup. You create a dot in the ChatGPT desktop app or in ChatGPT on desktop web. “You cannot currently create a dot on mobile, and dots are not supported on mobile web.” Afterwards you can talk to it in the mobile app “when mobile access is available.” [2]
• Timing. “Access may take several days to reach your account.” [2]

OpenAI’s pages describe texting in two ways. The announcement says “You can also message your dot in Slack and Teams”, with texting to follow. [3] The help page describes texting as “a limited beta” that is “limited to Pro users in the US and is not available in Business or Enterprise workspaces.” [2] We read both on the same day. Go by what your own account shows.

Which one for which job

“Not documented” means we read the pages and found no statement either way. “Plausibly” in the third row is our reading of that example. Whichever tool gives you an answer about a person, open the original before you rely on it; our piece on checking an AI citation shows how.

Where Kluro fits

Kluro publishes this article and makes a Mac app in the same broad area, a personal CRM. It overlaps with a dot in the way described above, and a dot does far more. Kluro does the jobs in the right-hand column: you ask about a person, place, topic or detail, it searches the conversations in the sources you have connected, and it brings back the relevant exchange for you to open. Saved memory stays encrypted on your Mac, and enabled hosted AI processes relevant content for answers and background understanding, so not everything happens on the device. A dot is not one of the sources Kluro connects to.

How we checked

On 30 September 2026 we read OpenAI’s announcement of dots, its post on safety, security and privacy, the dots help page, the dots privacy FAQ, the ChatGPT memory help page and the dots product page. We did not install, use or test a dot, and we ran no comparison between a dot and Kluro.

The two dots help pages show a relative “Updated” label and no fixed revision date. [2] [1] Read both again before you decide anything on a detail here.