Urgent.News

What's breaking now, across thousands of outlets.

AI

OpenAI launches always-on dots agents for Pro and Business

What dots are OpenAI has begun rolling out dots, its version of an always-on AI agent. According to a hands-on look from the Zapier blog , you talk to a dot through ChatGPT, but it works on its own Linux cloud computer with a real Chrome browser. That lets it keep going when your laptop is closed. It can also message you when it needs approval or something urgent turns up. The source says dots…

OpenAI has introduced "dots," its version of an always-on AI agent, available for ChatGPT Pro and Business Premium users. These agents operate independently on a Linux cloud computer with a real Chrome browser, enabling them to continue functioning even when the user's device is closed. Dots run on OpenAI's latest model, currently GPT-6 Astra, and are proficient in computer tasks.

They can handle emails, calendars, and Slack messages, flagging important information and preparing fixes for review. They can also analyze data as it arrives and draft overdue invoices. Users can access these features starting at $100 per month for ChatGPT Pro and $100 per user per month for ChatGPT Business Premium, billed annually.

Enterprise and Education customers can opt-in for a beta with custom pricing. Currently, free plans such as Go and Plus do not offer this functionality. OpenAI's testing has shown that dots can effectively manage various tasks, including detecting issues like bounced card payments and calendar conflicts. However, they are not yet ready to replace a user's primary calendar or inbox management system.

OpenAI acknowledges that while they have implemented safeguards to reduce the risk of prompt injection, these measures are not foolproof. Personal plans include default data training options, but users can disable this feature under settings. Business and Education plans do not have these defaults enabled.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in AI

When AI Starts Spending Our Money, How Do We Solve the Trust Problem?

In early 2026, daily active AI agents on-chain reached 250,000, up more than 400% year over year. These agents are no longer just chatbots.

  • AI agents on blockchain surged 400% in early 2026, reaching 250,000 daily active agents.
  • AI agents contributed $31 billion to DEX trading in 2025, accounting for 2% of total volume.

Classify data with an AI decision model

When a decision model beats a chat model for classification TL;DR For classification, use a decision model , not a chat model. A chat model writes text you then have to parse.

  • Decision models outperform chat models for classification tasks
  • Jev from TypeSafe provides probability-based answers for yes/no questions
  • Faster and more cost-effective labeling using Jev's calibrated probabilities

More from Saturday 17 October →