ChatGPT Dots use casesChatGPT Dots are OpenAI's always-on agents in ChatGPT, each with its own cloud computer, that watch your email and apps and message you when something needs action. Below are 6 real ChatGPT Dots use cases, each with the prompt, difficulty, setup time and what went wrong.
Dan Shipper tested OpenAI's Dots before launch. His Dot filtered incoming Slack requests and emails so he could triage without opening his inbox, and while he rescheduled a flight it warned that the new time clashed with a Slack meeting request he hadn't read.
A user's ChatGPT Dot joined a company's customer-service chat, introduced itself as their agent and negotiated a lower price. The store's bot negotiated back, then both bots asked their humans to confirm.
A user moved his bots' memory into an Obsidian vault so switching agents doesn't mean starting over. Grok Bot, Meta Muse, ChatGPT Dots and Hermes bots keep memory, logs and workflows in the same plain files he owns. Each bot reads the relevant files before answering and saves new information there (his health bot updates the health folder), and dated logs keep the history. A new agent gets pointed at the same vault with the same instructions.
A native, clearly labelled card in the directory results.
Become the featured sponsor →With his Google accounts connected to ChatGPT, a user's Dot messaged him unprompted about things that mattered: an assignment for his son's English class, a changed doggy-daycare pickup time, and a delayed American Airlines flight.
A founder told his ChatGPT Dot to apply to every startup program it could, fill everything in itself and authorize everything. He only filled a few fields it couldn't answer, and says he got $67,000 in startup credits in a day, with more software than he can use.
To let ChatGPT Dots communicate with Grok Bot, a user asked Grok Bot to set up a webhook, passed the link to his Dot, and that was it.