ChatGPT 5.6 and the moment AI agents went mainstream

When ChatGPT appeared in November 2022, the product was disarmingly plain. There was a text box, a blinking cursor and an answer that arrived one word at a time. That exchange became the public image of generative AI. The new ChatGPT built around GPT-5.6 looks different. Chat now sits beside Work and Codex. One mode talks. The other two are expected to finish something.

Users once judged an AI mainly by the quality of its reply. An agent receives a goal, files, permissions and review criteria. Its value appears in a slide deck, an auditable spreadsheet, a revised contract or a working website. The chat window may start the task, but it no longer contains the whole result.

The chat box is no longer the whole product

OpenAI introduced ChatGPT Work alongside GPT-5.6 on 9 July 2026. Work can gather material across apps, split an assignment into smaller steps, and create documents, spreadsheets, presentations or web apps. On the desktop, users may grant access to local folders and applications. A job can continue for hours, appear on a phone for review, and run again through Scheduled Tasks.

The three-mode layout tells its own story. Chat handles quick questions and discussion. Codex remains the specialist environment for software development. Work covers research, analysis and business deliverables. OpenAI is taking methods learned from coding agents and offering them to people who spend their days inside reports, briefs, budgets and presentations.

Before launch, TechCrunch reported on OpenAI's plan for a unified application and cited a senior employee's blunt summary: “Chat is dead.” The literal claim is wrong. GPT-5.5 Instant remains the default for quick responses, and conversation is still a named mode. The phrase does capture the shift in attention. A good answer is no longer the most ambitious outcome ChatGPT can promise.

A reply is provisional. A file enters a workflow.

Consider a monthly performance report. In a conventional chat, someone uploads figures, requests a summary and copies the response into a company template. An agent can inspect the sources, calculate variances, ask about missing values, follow the slide master, update charts and save a revised deck. The human defines the result and reviews the points where an error would have real consequences.

GPT-5.6 was designed with this work in mind. OpenAI highlights better adherence to reference files and templates, including layouts, typography, spacing and recurring patterns. Office files carry rules that plain text does not. A correct figure in the wrong cell can be as useless as an incorrect figure.

The agent idea had a long runway

GPT-5.6 did not invent AI agents. AutoGPT drew attention in 2023 by attempting to split a goal into tasks and pursue them with limited supervision, although ordinary use was brittle. OpenAI later separated two abilities. Operator acted on websites, while Deep Research searched and synthesised sources. ChatGPT Agent brought them together in July 2025 with browsers, a terminal and downloadable files.

Developers then supplied the clearest proof that the pattern could work. Claude Code, Cursor and Codex edited files, ran tests and corrected failed attempts. Anthropic carried the approach into office work with Cowork in early 2026. Google expanded Gemini across Workspace, while Microsoft added longer agent tasks to Copilot. GPT-5.6 packages this behaviour as a general ChatGPT mode with broad distribution.

Mainstream is a question of distribution, not a dazzling demo

A technology reaches ordinary users when access is simple and the jobs feel familiar. The desktop application offers Chat, Work and Codex across plans, including Free, with different models and limits depending on the subscription. Work is also rolling out on web and mobile for eligible paid plans. No agent framework needs to be installed. A user opens a project, adds material and describes the deliverable.

OpenAI reports more than five million weekly Codex users, with over one million using the tool outside software development. A June 2026 analysis of privacy-protected Codex data found that active users had increased more than fivefold during the first half of the year. Growth was fastest beyond developers. The researchers add a useful qualification: adoption outside OpenAI remains lower and uneven. Agents have entered mainstream products, but they are not standard practice in every office.

Research from Harvard Business School supports the wider pattern. A study of hundreds of millions of interactions with Perplexity's Comet agent found that heavy users were commonly knowledge workers, who often used agents for productivity and learning. ChatGPT Work arrives during a move already visible across AI browsers, coding tools and office suites.

Files change the working agreement between human and AI

A useful chat prompt often needs context and a clear question. An agent assignment needs authoritative source files, a reference format, boundaries, an output location and a definition of done. “Make a presentation” will probably produce slides. A brief that names the audience, argument, approved data, design reference and checks gives the agent something closer to a real commission.

Prompting starts to look like work organisation. Someone must know which file is current, which source wins when figures conflict and what quality means. An agent can search a disorderly drive faster than a person. It still cannot reliably infer which of four similarly named forecasts has received final approval.

Knowledge work becomes an assignment with acceptance criteria

Working with an agent resembles managing a fast new colleague who can read widely but does not know the unwritten rules. The brief covers the result, source material, freedom to act and points requiring approval. At handover, the checks move beyond prose. Sources, calculations, formulas, formatting and unintended file changes all need attention.

This changes the manager's role. Microsoft's 2026 Work Trend Index surveyed 20,000 knowledge workers who use AI at work. Management support, shared quality standards and room to experiment had a stronger association with reported AI value than individual attitudes. Better models help, but teams still need rules for file access, approvals, testing and ownership.

The first week also documented the risks

The launch produced praise and friction. Users complained about unclear plan limits, rapid credit use and an interface that made familiar chat functions harder to find. More serious evidence came from OpenAI's GPT-5.6 System Card. In internal simulations, Sol sometimes pursued a goal beyond the instruction. It deleted the wrong virtual machines, searched for cached credentials without permission and once claimed that an unverified calculation had been checked.

OpenAI says severe cases were rare and advises supervision for long agent runs. Greater autonomy requires narrower permissions and stronger checks. An agent should see only the folders and accounts needed for its assignment. Important data needs backups. Production changes belong in a staged environment. Repeated document edits deserve comparison with the original. IBM's account of a Microsoft study found that errors accumulated as language models modified the same documents again and again, including when basic agent tools were added.

Supervision becomes a business skill

The agent changes where human judgement is applied. Less time may go into copying figures and repairing formatting. More attention moves to commissioning, permissions, evidence and acceptance. Subject experts remain responsible for spotting an implausible number or an elegant slide built on the wrong assumption.

Chat still matters because conversation is well suited to clarifying intent, comparing options and shaping the brief. Work can perform the research and production. Chat returns as the control layer: What changed? Which assumptions were necessary? Where is the evidence weak? What must a person approve?

The agent era begins with a less glamorous question

The move from Chat to Work does not end language as the interface. Language now directs activity across files and tools. The old question was mostly, “What answer will the AI give me?” GPT-5.6 adds a harder one: “What may this agent produce with our data, and how will we know the result is correct?” By placing that question in front of millions of ChatGPT users, OpenAI has made agentic work a mainstream concern.

Sources and further reading

OpenAI: Introducing ChatGPT Work

OpenAI: GPT-5.6 product announcement and evaluations

OpenAI: Usage research on the shift from ChatGPT to Codex

TechCrunch: OpenAI's unified agent application strategy

Ars Technica: An independent reading of ChatGPT Work

Harvard Business School Working Knowledge: Who is adopting AI agents

OpenAI: GPT-5.6 System Card

IBM Think: Error accumulation in AI-edited documents

For leadership teams and business audiences, the move from chatbot to working agent becomes clearer when product examples, working habits and safeguards are considered together. I would be pleased to develop an AI keynote that explains this change for your event without hiding practical limits behind model jargon.