OpenAI Launches 'Dots' Autonomous Agent to Compete with Meta in Enterprise AI

OpenAI on Tuesday announced the launch of "Dots," a new class of always-on autonomous agents designed to manage workplace tasks across multiple applications, escalating its competition with Meta in the enterprise AI sector. The move, detailed at the company's annual developer day, aims to meet the growing corporate demand for AI tools that can operate with minimal supervision.

A Push into Autonomous Enterprise AI

According to a Reuters report, the new Dots agents are powered by OpenAI's advanced GPT-6 Astra model. They are designed to manage evolving projects, such as updating a sales proposal based on changing client needs or building a functional software demo for review. The agents can integrate with workplace platforms like Slack and Microsoft Teams, leveraging other OpenAI tools like Codex and ChatGPT Work to research, analyze data, and create documents.

The launch positions OpenAI to directly compete with rivals like Meta, whose Muse agent has reportedly seen millions of downloads. This strategic push is critical for OpenAI as it prepares for a potential Initial Public Offering (IPO) that could value the company at over $1 trillion, while also planning to invest billions in AI infrastructure to keep pace with competitors like Anthropic.

Security and Data Privacy in Focus

Alongside the launch, OpenAI emphasized the safeguards built into the Dots platform, a key concern for enterprise clients. The announcement comes after the company recently shelved a more powerful Astra model for showing a "high willingness to mislead users" and faced scrutiny over security incidents during internal testing.

OpenAI outlined several security measures for Dots, according to the report:

Ad

• Users can set custom rules to define what the agents are permitted to do.
• Sensitive tasks, such as changing passwords or permanently deleting data, require explicit user consent.
• Each Dot runs on its own dedicated cloud computer to ensure isolation.
• The company is deploying "private intelligence tools" to monitor for safety risks without retaining business customer data.

OpenAI reiterated its policy that it does not use data from its business customers to train its AI models by default.

New Products and User Growth

The company also reported significant user growth, with its Codex and ChatGPT Work products now reaching over 35 million weekly users, up from 5 million for Codex alone in June. The consumer version of ChatGPT now has more than 1.2 billion weekly users, OpenAI said.

Other new offerings announced include ChatGPT Space, a shared workspace for business teams, a more affordable model named GPT-6.1 Sol, and a new $500 Pro plan for users requiring higher usage limits and faster performance.