OpenAI is publicly conceding that its newest product didn't land smoothly. According to The Decoder, the company admitted it "didn't get everything quite right" with the launch of ChatGPT Work and is scrambling to fix problems with the user experience and costs.
ChatGPT Work went live on July 9, according to Outlook Business. Rather than a simple chatbot, it is described as an autonomous AI agent that merges ChatGPT with OpenAI's Codex coding tool. The idea is to have the software act less like a search box and more like a coworker: Outlook Business reports it can generate documents, spreadsheets, and websites on a user's behalf.
Under the hood, the product is powered by OpenAI's new GPT-5.6 model, per Outlook Business.
The tension in the story is the gap between ambition and execution. OpenAI is pitching ChatGPT Work as a tool that can take on real tasks automatically, but The Decoder reports the company is already acknowledging shortcomings in how the product feels to use and what it costs to run. The specific fixes OpenAI is pursuing were not detailed in the source material.
Why it matters: as OpenAI pushes AI from answering questions toward actually doing work, an early admission that the launch was rough is a reminder that "autonomous" AI coworkers still have real-world kinks to iron out before businesses can rely on them.