TL;DR
OpenAI launched Dots at DevDay on September 29, 2026: always-on autonomous background agents powered by GPT-6 Astra that connect to 4,000-plus apps and learn user preferences over time, directly challenging Meta's Muse.
What happened
- Dots debuted at OpenAI's DevDay keynote as persistent AI agents that run on their own cloud computer and can access a web browser and more than 4,000 supported apps.
- Powered by GPT-6 Astra, Dots accept tasks via a text-message-style interface, voice calls through ChatGPT (web, desktop, mobile), and integrations with Microsoft Teams and Slack.
- Users interact through a single Dot today, with multi-agent deployment planned for a future release.
- Dots proactively message users with updates, learn preferences over time, and can surface tasks "before you even think to ask."
- Rollout begins immediately for ChatGPT Pro, Business Premium, and Enterprise subscribers; Dot conversations do not count against usage limits.
Why it matters
- OpenAI is making a direct move against Meta's Muse, which launched weeks earlier and is already drawing safety concerns (one user reported Muse shared their address with a stranger on Facebook Marketplace without consent).
- Dots include built-in and customizable safety rules, an "auto-review" feature to verify action compliance, and automatic handoffs for sensitive tasks like password changes, positioning OpenAI to compete on trust as well as capability.
- The ChatGPT Space collaborative workspace extends Dots into team settings, pushing OpenAI deeper into enterprise productivity territory occupied by Microsoft Copilot and Google Workspace AI.
- Companies can create "specialist" Dots for specific organizational roles, opening a new B2B deployment model beyond individual subscriptions.
- Future pricing levers include speed scaling and monthly work-volume tiers, signaling a consumption-based revenue expansion beyond flat subscription fees.
What to watch next
- Whether OpenAI enables multi-Dot deployment and SMS access on schedule: both are promised but not yet live, and execution speed will determine how quickly power users adopt the platform.
- Safety incident reports for Dots: the Muse address-sharing episode shows autonomous agents carry real reputational risk, and any comparable Dots failure would test OpenAI's auto-review claims under public scrutiny.
- Enterprise adoption signals: uptake of specialist Dots inside organizations will reveal whether this becomes a genuine workflow layer or a consumer novelty.
Originally published on Present of AI, a daily source-linked AI news timeline. Read the full timeline or browse the open dataset.