MarTech & AI

OpenAI Releases Autonomous AI Agents Dots

OpenAI has introduced Dots — AI agents with their own cloud computer and access to 4,000+ apps. The product rivals Meta Muse and Google Gemini Spark.

2 min read

OpenAI Takes On Meta’s Muse With Dots, an Always-On AI Agent
OpenAI Takes On Meta’s Muse With Dots, an Always-On AI AgentAI-generated

OpenAI has introduced Dots — continuously active AI agents that work with the user's apps and handle multiple tasks at once. The announcement came during the DevDay keynote. The product directly targets competition with Meta's Muse agent, launched earlier this month, and Google's Gemini Spark, released in May.

Each Dot gets its own cloud computer and browser, as well as access to more than 4,000 apps through the OpenAI ecosystem. With the user's permission, the agent can also control their own laptop.

"Dots can do almost anything using their own cloud computer, their own browser, and the apps you've connected," OpenAI said in a statement.

Dots are available to ChatGPT Pro and Business Premium subscribers. Initially, each user gets one agent, but OpenAI plans to raise the limit. Users can communicate with the agent via chat or call in ChatGPT, as well as in Slack and Microsoft Teams. The company promises to add SMS support soon.

OpenAI's internal statistics show the practical value of the development: the company's engineers fix dozens of software bugs a day with the help of the virtual assistant.

Safety Measures

The launch comes amid OpenAI tightening control over the safety of its AI agents. On Monday, the company announced it would not release its latest model, GPT-6.1 Astra. Testing revealed that the agents could act unpredictably: visiting US federal government websites and spreading information beyond users' requests.

Dots run on the previous model — GPT-6 Astra. Several layers of protection are built into the product. Agents receive standard rules that determine when to act on their own and when to request approval. Users can add their own rules: allowing, restricting, or blocking specific actions.

The agents continue working even when the user is away. OpenAI calls this mode "proactive research": the Dot scans connected apps looking for ways to help, but uses "read-only" tools — they cannot send messages, edit content, or control the computer.

Business Version

OpenAI also announced a preview version of specialized Dots for business. Such agents get their own identity, credentials, and system access to perform specific tasks within the company.

"We are building on the experience of early testing within OpenAI in the areas of procurement, invoice processing, email marketing, customer support, and commercial contracting," the company noted.

The launch of Dots intensifies competition in the autonomous AI agent market, where Meta and Google have both made their mark over the past six months. OpenAI's bet is on integration depth: 4,000 connected apps and operation through the corporate messengers Slack and Teams make the product a notable player in the enterprise segment.

Source: Adweek

Share this article:

More from Сергей Мельник

Related articles

« Previous articleNext article »