OpenAI launches Dots, always-on AI agents, for $100-a-month ChatGPT Pro users
Announced at DevDay 2026, Dots work in the background, connect to your apps and handle multistep tasks, asking permission before sensitive actions.
By Eric Powlesland 3 min read
OpenAI has introduced Dots, a new kind of AI assistant that keeps working on a task after you stop typing. The company announced them on Tuesday at its DevDay 2026 event in San Francisco, as Wired reported. Each Dot appears as a small, customizable blob that you can hand a multistep project to.
The difference from a regular chatbot is the "always-on" design. Instead of answering one question and waiting, a Dot keeps searching the web and working toward whatever goal you give it. It also draws context from apps you connect and learns your preferences over time. Dots run on OpenAI's GPT-6 Astra model.
In one OpenAI demo, a Dot noticed on a user's calendar that they would be working through dinner. It then sent two GrubHub delivery options with prices, and the user replied with a choice and a delivery time. In another demo, a user worked with a Dot to launch a new website.
You can message a Dot through ChatGPT, Slack or Microsoft Teams, and it carries the same context across all three. Pro subscribers can also join a waitlist to text their Dot through iMessage or RCS messaging on Android.
The rollout begins with ChatGPT Pro, the $100-a-month tier. OpenAI often starts new features there before opening them to more users. For now, each user controls one Dot, though Wired says the ability to run several at once is expected later.
OpenAI is also aiming Dots at larger organizations. On stage, CEO Sam Altman announced "specialist Dots" for enterprise customers, tuned for work such as accounting, email marketing and legal analysis. Pricing and availability details for those versions were not included in the report.
What it means for your business
For most small businesses, the immediate impact is limited. Dots are only available to Pro subscribers today, and the enterprise specialist versions do not yet have published details. But the direction is clear: AI tools are moving from answering questions to doing work in the background, inside the apps your team already uses.
If your staff live in Slack or Teams, this is worth paying attention to. An assistant that can watch a calendar, gather options and come back with a recommendation could save time on routine coordination. That said, the demos shown were simple, and how well Dots handle messier, real-world business tasks is not yet known.
The bigger issue is trust. An agent that reads your connected apps and acts for you has access to a lot of information. Dots are designed to ask for explicit approval before sensitive steps like installing software or changing a password, and a Custom Rules tool lets you set limits and require permission for certain tasks. Even so, Wired cautions that agents could share information you did not intend to share, or be tricked by a bad actor into taking actions you did not want.
One setting deserves a check: if your OpenAI account allows your data to be used for model training, that setting also applies to your conversations with Dots.
If you or someone on your team tries Dots, some sensible starting points:
- Connect only the apps the task actually needs.
- Set Custom Rules before assigning real work.
- Review your account's data training setting.
- Keep client and financial data out of early experiments.
What to watch
Dots arrive as personal agents are drawing wider interest. Meta's Muse agent recently climbed to the top of smartphone download charts, and Dots look like OpenAI's direct response. Watch for when Dots reach lower-priced ChatGPT plans, when multiple agents per user become available, and what the enterprise specialist Dots cost. Wired notes this is a developing story, so more details may follow.
