OpenAI CEO Sam Altman introduced Dots during the company’s OpenAI DevDay on Tuesday. Dots is designed to handle tasks for users over long periods instead of only answering questions when asked.
The AI agent can connect to about 4,000 apps through ChatGPT. This means users can give Dots ongoing tasks involving the different tools and services they already use. For example, a user could tell Dots to monitor an ongoing project, keep them updated about its progress, or add items to their schedule. The idea is to let the AI handle routine work in the background.
How OpenAI Dots works
OpenAI says Dots is powered by its Astra model and can recognize events that need action. For example, if a software bug alert appears in Slack, Dots can notice the alert and start looking into the problem.
OpenAI gave an example involving an early tester who had not submitted an invoice for a publication. Dots noticed the missing invoice, prepared it and submitted it after the tester approved the action.
Users will be able to hand tasks over to their Dot through Slack. This could help users deal with the large number of messages and notifications they receive during the workday. OpenAI says users will start with a single Dot. They can give it a name and customize how it works for them.
Also read: Anthropic IPO: What investors should know as leaked filing reveals huge AI costs and risks
Dots targets business users
The company says it could eventually allow people to have entire teams of Dots working together. This points to OpenAI’s broader push toward AI agents that can manage multiple tasks instead of simply responding to individual prompts. The new AI agent will roll out to Pro, Business Premium and Enterprise users in markets where it is available. This puts Dots directly into the workplace and enterprise AI market.
OpenAI’s launch comes weeks after Meta introduced its Muse AI agent. Muse quickly reached the top of both Apple’s App Store and Google Play Store, giving Meta a strong early result for its consumer-focused AI product, according to Yahoo Finance.
Dots vs Meta Muse
Meta is marketing Muse mainly toward consumers, while OpenAI is positioning Dots for the enterprise and workplace market. This means the two companies are using AI agents for different parts of the market. Experience with Meta’s Muse found that the agent can sometimes make mistakes but generally works well for average users. It can be useful for tasks such as comparing products and getting quick answers to complicated questions. Yahoo Finance’s reporter said this was based on their experience using Muse.
Also read: Anthropic IPO: Can its $2 trillion valuation survive a $42 billion loss and massive AI spending?
OpenAI, Anthropic, Meta and Google have all faced situations in which their AI agents reportedly went beyond their intended tasks, including hacking activity involving organizations and governments over the past year.
AI agent safety concerns
Amodei has urged AI companies to slow the development of frontier AI models after concerns were raised about the risks from increasingly powerful systems. Sam Altman has said he agrees with the need for caution, according to Yahoo Finance.
The company unveiled its latest AI model, GPT-6.1 Sol, during the same event. According to OpenAI, GPT-6.1 Sol is nearly as powerful as its top-end Astra model for coding, computer use and professional work, but costs significantly less.
AI companies are facing greater pressure to make their models cheaper because businesses are looking for ways to reduce how much they spend on AI. This makes pricing an increasingly important part of the competition between AI labs.
OpenAI’s annualized revenue was reportedly around $70 billion at the beginning of the third quarter, according to Axios. OpenAI is reportedly preparing to launch an initial public offering early next year, after delaying plans that had reportedly been considered for this fall.
Dots shows OpenAI’s push beyond chatbots toward AI systems that can monitor work, connect to workplace apps, identify tasks and take actions with user approval. Meta’s Muse is competing in the consumer market, while OpenAI is using Dots to target businesses and professional users.