OpenAI used its Dev Day event to launch something new inside ChatGPT: Dots, a type of AI agent meant to work on your behalf in the background, even while you're not logged in. If you've mostly used ChatGPT as a chat window you type questions into, this is a meaningfully different way of using it.
What is a ChatGPT Dot, in plain terms?
A Dot is an AI agent, built by OpenAI, that runs on its own, separate from your device, and keeps working on a goal you've given it over time rather than waiting for you to ask a question each time. You talk to it the same way you'd talk to a person: message it or call it inside ChatGPT (desktop, web, or mobile), Slack, or Microsoft Teams, tell it what you want, and it keeps working, checks in when it needs a decision from you, and learns your preferences as it goes.
That's the core difference from regular ChatGPT. A normal ChatGPT conversation only does something while you're actively typing to it. A Dot can keep working on a task across hours or days without you sitting there.
What can it actually do?
OpenAI's own examples give a sense of the range: a developer assigning a Dot to turn user feedback into tested code fixes, a marketer having one revise launch materials as a project's scope changes, a scientist letting one update an analysis whenever new data comes in, a sales team using one to help build out proposals. Dots can also connect to over 4,000 outside apps through OpenAI's plugin ecosystem, which is what lets them actually take action in other tools instead of just talking about it.
There are guardrails built in. Sensitive actions, like changing a password, always require your approval first. You can set your own custom rules for what a Dot is allowed to do without checking in, and there's an Activity View where you can see exactly what it's done.
Who can get it, and what does it cost?
Dots are rolling out now to ChatGPT Pro and Business Premium subscribers in eligible markets, with one Dot included at no extra cost beyond your existing subscription. Enterprise customers can get access to a beta version through their workspace admin. OpenAI has said it plans to offer additional Dots and other scaling options later on.
Worth knowing if privacy is a factor for your organization: OpenAI's existing data policies carry over, and Business and Enterprise conversations aren't used to train OpenAI's models by default.
So what does this mean for your organization?
If your team is already using ChatGPT day to day (for drafting appeals, summarizing meeting notes, brainstorming campaign copy), Dots represent a step toward a different kind of use case: a standing assistant assigned to one ongoing job rather than a tool you open for a single task. A grant deadline tracker that checks in on its own, or a Dot assigned to keep a set of donor acknowledgment letters current as new gifts come in, are the kind of jobs this is aimed at.
Two things worth weighing before jumping in. First, the free tier of ChatGPT doesn't include this; it currently requires a Pro or Business Premium plan, so it's worth checking that cost against the actual time it would save your team. Second, since a Dot can connect to outside apps and take real actions, it's worth reading through the approval and permission settings carefully before connecting it to anything that touches donor data, financial systems, or anything else sensitive to your organization.
This kind of always-on AI agent is still new enough that best practices for using it well, especially for a small team without a dedicated IT function, are still being worked out in real time. If you'd like help thinking through whether a tool like this fits into your organization's workflow, and how to set it up safely if it does, reach out to Tapp Network. We're already testing tools like this in our own operations and can help you sort the genuinely useful parts from the hype.
.jpeg)
