OpenAI Launches 'Dots' Personal Agent, Joining Meta in Race for Always-On AI Assistants

OpenAI unveiled Dots, an "always-on" personal AI agent, at its DevDay conference in San Francisco on September 30, 2026, positioning the company directly against Meta's Muse, which launched three weeks earlier and has driven Meta's stock up 29% in September — its best month since 2013.

CEO Sam Altman described Dots as "remarkably capable" and "built to handle really anything you can think of," according to OpenAI's DevDay keynote. The product is designed to complete tasks such as managing Slack and Microsoft Teams workflows, though OpenAI has not disclosed specific benchmark results or technical specifications for the underlying system.

Access to Dots requires OpenAI's Pro plan, priced at $100 per month or higher — a notably steeper entry point than Meta's approach. Muse is free with a usage cap and converts to a paid tier starting at $20 per month. Neither company has disclosed a context window size, parameter count, or training cutoff date for the systems powering these agents, since both are framed as product layers rather than standalone foundation models.

A crowded field

OpenAI and Meta are not alone. Google markets Gemini Spark as "your 24/7 personal AI agent," and Apple has redesigned Siri, calling it "a profoundly more capable and personal assistant." Startup Instinct raised $1 billion from investors including Sequoia at a $10 billion valuation this week, and rival startup Town is reportedly in talks at a $1 billion valuation.

Town CEO Jean-Denis Grèze told CNBC that OpenAI's Dots launch "continues to validate the category and the need for work assistants," adding that Town has focused on this space "from the very beginning."

Homebase CEO John Waldmann, who uses various agent tools, said the near-simultaneous launches reflect where the underlying technology has landed rather than coordinated strategy: "It's probably no surprise that it's all happening around the same time." He was more skeptical about staying power, questioning whether agents will become part of users' daily habits or simply get tried once and abandoned.

Thomas Randall, a research director at Info-Tech Research Group, said the market will likely be decided less by raw model capability and more by distribution: "Once leading systems are broadly capable, the winner will likely be determined by which agent is the easiest to adopt, the most connected, and the hardest to leave."

Context: safety debate in the background

The Dots launch came the same day OpenAI president Greg Brockman attended a Washington lunch hosted by President Trump focused on AI safety — over two weeks after Anthropic CEO Dario Amodei published an essay urging AI labs to slow frontier model development. The agent product push offers labs a more optimistic narrative than the ongoing safety scrutiny surrounding frontier model releases.

What this means

Dots is not a new foundation model — it's a product wrapper for existing OpenAI capabilities, gated behind a premium subscription tier. That makes this less a technical leap and more a distribution and monetization battle. With Google, Apple, Meta, OpenAI, and well-funded startups like Instinct and Town all converging on nearly identical pitches — an agent that manages your calendar, emails, and tasks — differentiation will hinge on integration depth and retention, not raw model scores. OpenAI's $100/month price floor is a bet that professional users will pay a premium for agentic reliability; Meta's freemium approach bets on volume. Whether either strategy converts trial downloads into daily habitual use, as Waldmann noted, remains the open question that will decide which company actually wins the category.