OpenAI Unveils 'Dots': Your New Always-On AI Agent Is Here!

OpenAI has just rolled out its latest innovation: "Dots," new always-on AI agents designed to transform how we interact with technology. Announced at the company's DevDay 2026 event in San Francisco, these agents are depicted as customizable, cute blobs that can be assigned multi-step projects and complete tasks on your behalf. Powered by OpenAI's advanced GPT-6 Astra model, Dots stand out with their "always-on" design.

Unlike traditional single-turn chatbots, these intelligent agents continuously crawl the web, pull context from your connected apps, and learn your preferences over time to proactively assist you. Imagine an AI that not only remembers your habits but actively works to simplify your day! OpenAI shared compelling demos showcasing Dots in action.

In one instance, a user's Dot noticed a dinner conflict on their calendar and proactively messaged two GrubHub food delivery options, including pricing.

Another demonstration showed a user collaborating with a Dot to launch a new website.

These examples highlight the agents' ability to handle practical, real-world scenarios.

Accessibility for Dots is broad, allowing users to message them through ChatGPT, Slack, and Microsoft Teams. For ChatGPT Pro subscribers, who pay $100 a month, there's an exclusive waitlist to text with their Dots via iMessage or RCS messaging on Android. OpenAI typically introduces new features to its higher subscription tiers first, with a broader rollout planned for later.

Beyond personal use, OpenAI is also targeting the enterprise sector. CEO Sam Altman revealed "specialist Dots" for business clients during DevDay, tailored for specific professional tasks like accounting, email marketing, and legal analysis. This move further integrates AI agent coworkers into professional workflows, signaling a significant shift in enterprise operations.

While the prospect of an always-on AI assistant is exciting, OpenAI has emphasized safety and user control.

Dots are designed to seek explicit approval before performing sensitive actions like installing software or changing passwords.

Users can also set explicit boundaries using a "Custom Rules" tool.

However, connecting personal data to these tools always warrants caution, as users should be mindful of potential security and privacy risks. It's also important to note that if your OpenAI account settings allow model training on your data, this will extend to your interactions with Dots. The launch of Dots enters a rapidly evolving landscape for personal agents.

With Meta's Muse agent recently gaining significant traction, OpenAI's Dots are poised to intensify the competition. This innovation marks a new chapter in personalized AI, promising to make our digital lives more seamless and intelligent. #OpenAIDots #AIagents #TechInnovation #DevDay2026