MilikMilik

ChatGPT Work Turns AI Into a Colleague for Long-Running Projects

ChatGPT Work Turns AI Into a Colleague for Long-Running Projects
Interest|High-Quality Software

From Chatbot to Colleague: What ChatGPT Work Really Is

ChatGPT Work is an autonomous AI agent inside ChatGPT, powered by GPT-5.6, that can stay on multi-step projects for hours, moving across your connected apps, local files and web services to gather information and produce finished work products like documents, spreadsheets, presentations and web applications without needing constant human supervision. This is not another prompt-and-reply mode; it is OpenAI’s bid to make ChatGPT behave like a digital colleague who owns a workflow from brief to deliverable. The company has launched ChatGPT Work alongside the GPT-5.6 model family, which now runs both Work and the integrated Codex coding tools. In effect, OpenAI is declaring that chat time is over. The strategic bet is that knowledge workers want “an agent that does the work, not just one that writes code,” backed by more than five million weekly Codex users.

ChatGPT Work Turns AI Into a Colleague for Long-Running Projects

Hours-Long AI Multi-Step Automation, Not One-Off Answers

The key shift with the ChatGPT Work agent is persistence: it stays on complex projects for hours, breaking them into smaller steps and working through them independently. Users can give it goals like analysing a monthly budget, preparing for a sales meeting or turning research into a campaign brief, then let the agent divide the assignment, gather data and output spreadsheets, presentations, documents and interactive web apps. This AI multi-step automation is a clear break from traditional chatbots that stop at a single response. Work gathers information across connected apps and workflows to build finished materials instead of drafts. According to OpenAI, “users can assign an entire workflow through one request,” meaning the agent can carry context from early research through to campaign assets for different markets without losing the thread. That is the behaviour of a junior analyst, not a search box.

ChatGPT Work Turns AI Into a Colleague for Long-Running Projects

Connected Apps, Scheduled Tasks and the New Autonomous Workflow Tools Stack

ChatGPT Work is opinionated about where real work lives: inside calendars, CRMs, chat apps and shared drives, not in a vacuum. It connects to services including Slack, Microsoft Teams, Google Drive, SharePoint, Gmail, calendars, customer relationship management platforms and project trackers, pulling from information already spread across an organisation. On desktop, it gains direct access to local files and installed applications plus a built-in browser that can gather live website data and work with files in Google Workspace and Microsoft 365. Plugins act as the wiring between the agent and these tools, while Scheduled Tasks turn it into a true autonomous workflow tool: reviewing new messages, refreshing meeting documents, checking dashboards, and updating presentations whenever new feedback arrives. Users still choose what it can access and which actions need explicit approval, but the day-to-day drudgery starts to move off their plate.

ChatGPT Work Turns AI Into a Colleague for Long-Running Projects

GPT-5.6 and the Battle to Own Long-Running Enterprise AI Agents

Underneath Work sits GPT-5.6, a model family built explicitly to manage longer workflows and reason through multi-step tasks while matching existing templates and reference files. The flagship Sol model costs USD 5 (approx. RM23) per million input tokens and USD 30 (approx. RM138) per million output tokens via API, with Terra at USD 2.50 (approx. RM12) input and USD 15 (approx. RM69) output and Luna at USD 1 (approx. RM5) input and USD 6 (approx. RM28) output. On Agents’ Last Exam, a benchmark for long-running professional workflows, Sol scores 53.6, 13.1 points ahead of rival model Claude Fable 5. This pricing and performance story matters: OpenAI is signalling that multi-hour, multi-app agents are now a first-class product line. The company will retire GPT-5.4 on July 23, keeping GPT-5.5 models available as the 5.6 lineup rolls out over 24 hours, a tidy transition from short chats to durable AI co-workers.

What It Means for Knowledge Workers: From Requests to Results

The practical impact is immediate: ChatGPT Work is rolling out on web and mobile to Pro, Enterprise and Edu users, with Plus and Business plans following within days. Free and Go users are excluded, which is telling; this is built for people whose time is money. Early testers show the change in posture. One head of enterprise marketing used Work to build a repeatable system for reviewing thousands of leads monthly, tracing CRM and email touchpoints, flagging broken follow-ups and generating a weekly executive dashboard that surfaced seven figures in missed pipeline. That is not a chatbot “helping brainstorm email copy.” It is an agent running a living workflow. As OpenAI puts it, ChatGPT is now “a partner for your most ambitious work” rather than a place to ask quick questions. The conclusion is hard to dodge: for knowledge workers, the interface of work is shifting from apps to agents.

Milik earns a commission when you shop through our links, at no extra cost to you. This article was generated with AI from published sources and product data.

You May Also Like

Comments
Say something...
No comments yet. Be the first to share your thoughts!