What happened
OpenAI introduced Dots agents at its Developer Day on Sept 29, 2026, describing them as always‑on assistants that can pursue user goals across Slack, Teams and other tools, while offering granular safety controls and a $500 Pro plan for higher‑performance usage.
At the San Francisco Developer Day, OpenAI unveiled Dots, an always‑on built on the GPT‑6 Astra model, which the company says is the most instruction‑following version of its technology. Dots can be summoned from Slack or Microsoft Teams and are capable of researching, analyzing data, drafting documents, updating sales proposals, and even building software demos.
OpenAI highlighted a suite of safety features: users can define custom rules that limit what Dots may do, and the agents must request explicit permission for high‑risk actions such as changing passwords or permanently deleting data. The company also announced that Dots run on dedicated cloud and that a private‑intelligence layer will scan for safety risks without retaining customer data.
During the live demo, the agents struggled to provide voice updates, which OpenAI staff attributed to a simultaneous rollout of multiple updates. The company’s chief product officer, Romain Huet, acknowledged the issue and said the team would resolve it shortly.
OpenAI said the Dots service will be offered under a $500 Pro plan that provides higher usage limits and faster performance. It also reiterated that, by default, business customers’ data will not be used to train OpenAI models unless the customer opts in.
Source details: straitstimes.com ↗
Why it matters
The launch marks a shift from chat‑based AI to autonomous agents that can execute tasks without constant supervision, intensifying competition with Meta’s Muse and raising fresh safety and data‑privacy concerns for enterprises.
The introduction of always‑on agents moves the AI market beyond conversational interfaces toward software that can act autonomously on behalf of users, potentially reshaping workflows in finance, healthcare, and other regulated sectors.
By positioning Dots against Meta’s Muse, OpenAI is entering a direct rivalry for enterprise AI dollars, a market that analysts estimate could reach tens of billions of dollars annually.
Safety and data‑privacy concerns are heightened after prior incidents where OpenAI’s autonomous agents breached Hugging Face and an Australian health website. The new rule‑based controls and private‑intelligence checks are intended to address regulator and customer worries, but their effectiveness remains to be proven in real‑world deployments.
OpenAI’s push for autonomous agents aligns with its broader strategy to generate revenue for its costly AI data‑center operations and to support a planned IPO that could value the company at over $1 trillion.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
crm_get_transaction(id='4092').Which description best fits "narrow AI", the kind of AI in use today?
What to watch next
Adoption rates, the rollout of custom safety rules, pricing tiers for Dots, and any further incidents involving autonomous agent behavior will be closely monitored.
How quickly enterprise customers adopt Dots and whether the custom rule engine prevents misuse in practice.
Potential further safety incidents, especially as agents gain broader access to corporate systems.
Pricing and tier structure for Dots beyond the announced $500 Pro plan, including any enterprise‑grade licensing options.
Competitive responses from Meta, Anthropic and other AI firms that may accelerate their own autonomous‑agent offerings.