AI

OpenAI has introduced ChatGPT Work, a new agent designed to help users complete longer, more involved projects inside ChatGPT. Announced July 9 and now rolling out to eligible accounts, Work can research information, use connected apps and files, create finished materials and continue working through multi-step assignments that would ordinarily require several separate prompts. The simplest way to understand Work is as the “get things done” side of ChatGPT. A standard conversation may help someone think through a problem or produce an individual answer. Work can take responsibility for a broader workflow—gathering information, planning the steps and turning a user’s goal into a finished, reviewable result. Powered by OpenAI’s new GPT‑5.6 model and technology developed for Codex, Work can create presentations, spreadsheets, documents, reports, websites and other creative assets. It can also assist with research, content production, data analysis and software projects, while breaking especially complex assignments into smaller tasks that can be completed independently. Scheduled Tasks extend those capabilities beyond a single session. Users can arrange for Work to complete a task once, repeat it on a schedule, respond to a future trigger or monitor for changes. A digital publisher, for example, could schedule a morning news search that identifies promising stories, organizes them by editorial channel and prepares article drafts and supporting materials for review. Users remain involved throughout the process. They can follow Work’s progress, answer questions, change direction and approve important actions before they occur. That combination of sustained execution and human oversight signals a meaningful evolution for ChatGPT: from an AI that primarily discusses the work to one increasingly capable of helping complete it.