OpenAI Says Over a Quarter of U.S. Workers Now Use ChatGPT on the Job

Over a quarter of U.S. workers now report using ChatGPT for work, and 45% of those with postgraduate degrees do the same, according to a new OpenAI report on how businesses use GPT-5 and ChatGPT.

The report, which OpenAI published as "Inside GPT-5 for Work: How Businesses Use GPT-5," combines third-party industry studies with OpenAI's own analysis of ChatGPT and ChatGPT Enterprise usage. OpenAI states that all its analyses were performed on anonymized or aggregated usage data using automated content classifiers, and that the company did not review any user or customer content. Prompts cited in the report are synthetic examples written for illustration, not real user queries.

The stakes are straightforward. ChatGPT has grown from a product aimed at AI researchers into one of the world's most visited websites, with over 700 million weekly active users. Launched in November 2022, it reached 100 million weekly active users within months. How deeply a general-purpose AI assistant embeds itself in everyday work across industries will shape enterprise software spending, hiring, and productivity for years.

Bottom-up adoption, not top-down rollout

OpenAI argues ChatGPT broke the traditional enterprise software pattern of big upfront costs, long rollouts, and slow adoption. Employees brought it into their jobs from their personal lives without formal training or onboarding.

"This grassroots pattern has made it the fastest adopted enterprise technology in recent history," the report states.

The adoption path mirrors earlier consumer-to-enterprise software shifts, driven heavily by younger employees. OpenAI points to ChatGPT's high penetration among workers under 30 and frequent, often daily, use.

Which industries lead, which lag

IT and finance show the highest-than-expected adoption rates, which OpenAI attributes to the tool's strengths in coding, analysis, and information-heavy work. Manufacturing adoption signals a broader digital transformation: factories using AI for process automation, predictive maintenance, and supply chain optimization, with early industrial AI investments paving the way for use among engineers, analysts, and operations managers.

Retail, construction, transportation, wholesale trade, and agriculture all show significantly lower adoption. OpenAI notes this tracks with those sectors' smaller share of knowledge workers.

Healthcare is a special case. Despite being one of the largest and most data-intensive sectors, adoption has been slower, which OpenAI attributes to strict privacy and compliance rules and risk-averse organizational cultures. The report sees growth in targeted areas like clinical documentation and administrative workflows, suggesting healthcare "could soon become a hotbed of AI adoption."

What people actually use it for

In the first three months of usage, four categories dominate: writing, research, programming, and analysis. Together they account for the majority of messages sent.

Technical teams are among the heaviest users. Analytics, engineering, and IT roles make up a large share of early usage. Programming is the top task for engineering roles, but users also request substantial research and documentation help — suggesting ChatGPT is used nearly as much for planning as for coding. IT teams lean on research and troubleshooting, often treating ChatGPT as an information resource before moving into automation.

Go-to-market roles — marketing, communications, sales, and customer experience — rely on the tool primarily for writing, research, creative ideation, and media generation.

The report stresses a consistent pattern: "AI is augmenting expertise, not replacing it." Engineers iterate on prompts to debug code and generate unit tests. Analysts use chain-of-thought prompting to clean and interpret datasets. Customer support teams draft brand-aligned responses.

Coding is spreading beyond engineering. Designers use ChatGPT for coding at a much higher rate than finance and sales, likely for front-end prototyping and snippet help. Project managers combine writing, media generation, coding, and data analysis. Product, operations, marketing, finance, and HR all use ChatGPT for coding to some extent.

Design teams stand out for media generation, relying on it 2–4x more than other groups — evidence, OpenAI argues, of an emerging role for ChatGPT beyond text.

Measured productivity gains

OpenAI cites a study by Boston University and BCG examining ChatGPT's impact on the technical competency of BCG consultants. Consultants armed with and trained on ChatGPT scored 49, 20, and 18 percentage points higher than the control group on three technical tasks, and performed close to the level of real BCG data scientists on two of the three tasks.

A separate six-month randomized field experiment across thousands of knowledge workers found that access to AI cut weekly email time by 31%. Another study of software developers found AI coding tools enabled them to spend more time coding, more time on exploratory work, and less time on project management.

OpenAI's internal benchmarks show "meaningful increases in productivity," driven by employees who use the tool to write and communicate faster, research more effectively, and reduce effort on repetitive tasks. The company acknowledges most organizations are still in the early stages of adoption.

Advanced features remain underused

Most departments rely on core ChatGPT tools: search, data analysis, file uploads, retrieval, and canvas. More advanced features — reasoning models, deep research, projects, and custom instructions — see higher use mainly among power users, including R&D teams.

Technical functions are the exception. Analytics, engineering, IT, and research roles are much heavier users of advanced capabilities, reflecting work that demands multi-step reasoning, large-scale data synthesis, or complex problem-solving. IT teams favor retrieval and search for configuration and policy questions; engineering teams show stronger use of GPTs, programming tools, and data analysis.

OpenAI identifies this gap as an opportunity, noting barriers may include discoverability, awareness of use cases, or the setup required. GPT-5 addresses part of the problem with a real-time router that automatically decides which advanced features and tools to use based on conversation type, complexity, tool needs, and explicit intent.

Usage intensity is growing alongside user counts. Certain power-user segments of ChatGPT Pro subscribers send upwards of 200 messages per day, and usage has evolved from simple Q&A to coding, data analysis, and agentic workflows.

From tool to platform

OpenAI frames the next phase as a shift from personal productivity to workflow platform — what the report calls an "operating system for daily work." New capabilities, from autonomous agents to advanced coding support to decision-assist tools, are expanding ChatGPT's role. Executives use it to shape strategy, engineers to design and debug systems, and customer support agents to evaluate complex solutions.

Looking ahead, the report predicts employees will spend less time performing tasks and more time supervising and shaping AI output, with individuals taking on tasks once spread across multiple departments — a product manager, for example, analyzing customer feedback, testing a feature, and drafting the legal and marketing content to bring it to market.

OpenAI draws a historical parallel: electricity reshaped factory work, the internet redefined commerce and communication, and AI is setting the stage for the next leap. The enterprises that adapt quickly, the report argues, will capture the earliest and largest gains — faster decision cycles, productivity breakthroughs, and new opportunities across every function.