OpenAI Unveils Always-On Dots Agents and ChatGPT Team Tools

According to Slashdot, which cited Bloomberg, Dots can research, draft documents, write software and handle other tasks, and each user will initially have access to a single 'dot' that acts as an assistant. Over time, OpenAI expects users to manage multiple dots. OpenAI's announcement said the agents are powered by GPT-6 Astra, have their own cloud computer, learn from feedback and can work toward a user's goals 24/7. Through an ecosystem of plugins, they can connect to more than 4,000 apps, and users can reach them through ChatGPT, Slack or Teams, including by voice call. 'The more you work together, the more your dot learns your preferences, how you think, and what good looks like to you,' OpenAI said. Slashdot also reported that OpenAI announced GPT-6.1 Astra; the Dots announcement text referred to GPT-6 Astra.

For Pro, Business and Enterprise plans, OpenAI introduced ChatGPT Space, described in a blog post as 'a new home for your team to collaborate with AI to get work done.' Users can work with ChatGPT, teammates and Dots, and can try Space today on the desktop app and web, ZDNet reported, with mobile access coming later. Pages and collaborative slides, also announced Tuesday, are 'built for human and agent collaboration,' OpenAI said. They let users create documents and presentations with generative abilities, including on-page research, graphics and images. Users can design slides conversationally or import templates, and multiple users and agents can edit simultaneously. Presentations can be given directly from ChatGPT or exported to PowerPoint or Google without disrupting formatting, according to ZDNet. Pages are available now, while slides are expected in the next few weeks.

OpenAI CEO Sam Altman discussed Space, Pages, slides and Dots as pieces of a new system during the keynote but did not frame them as a Google Workspace competitor, ZDNet reported. An OpenAI representative told ZDNet that Space is built for people and AI to work on the same content together, while Pages are a ChatGPT-native format that AI can read and create and teammates can edit in real time. 'Today, people develop ideas in ChatGPT, move the work elsewhere to share it, then bring feedback back to ChatGPT,' the representative said. 'Space brings that process together so people can create, share, and keep working with their team and ChatGPT in one place.'

OpenAI also announced Private Intelligence, a data access approval layer intended to give enterprise customers greater confidence that their data is protected. The company said Zero Data Retention with Private Safety Processing enables automated safety reviews without giving OpenAI personnel access to the underlying content. A preview of Private Inference, planned for this fall, combines confidential computing with strict, verifiable controls, OpenAI said, without providing more detail on how Private Intelligence and Private Inference differ, according to ZDNet. The name is reminiscent of Apple's Private Cloud Compute.

ZDNet noted that the releases resemble Anthropic's recent launch of Claude Docs as part of its Claude Cowork consolidation, which also absorbed Claude Design and Slides. OpenAI launched ChatGPT Work in July as an answer to Claude Cowork, and Anthropic has since folded Cowork back into Claude. Last year, OpenAI was reported to be exploring features that would turn ChatGPT into more of a productivity suite, including collaborative document editing similar to Google Docs, but it has since launched and sunset its consumer video app Sora while Anthropic gained enterprise revenue.

Separately, Okta Inc. Chief Revenue Officer Jon Addison said in an interview with theCUBE at Okta's Oktane event that AI agents are pushing enterprises to modernize and consolidate identity and access management. According to SiliconANGLE, many enterprises built their identity stacks piece by piece and still spend heavily to keep point solutions connected. Agents raise the stakes on that sprawl, Addison said, because point solutions create vulnerabilities. 'They're getting their fundamentals right,' he said, describing the work as housekeeping ahead of the agentic era. Some organizations are agreeing on reference architectures before deploying agents, while others already have shadow agents in production and need urgent access management controls, he said. Addison described the technology as offering efficiency and augmentation for human workers, but warned that autonomous agents will eventually scale in ways that require different thinking.

Editor's Summary

OpenAI's DevDay announcements put always-on Dots agents and ChatGPT-based collaboration tools at the center of its workplace strategy, while adding enterprise data controls. Okta's chief revenue officer said the same shift is exposing fragmented identity systems and forcing consolidation. The combined moves show AI vendors competing to embed agents in daily work, with identity and security controls becoming a prerequisite for broader deployment.