AI Revolution in Business: OpenAI Predicts Massive Adoption by 2025

In early December 2025, OpenAI released its annual report on the state of AI in business, which includes some staggering statistics. The California-based company now serves over a million professional clients, with usage indicators showing a significant boom in adoption. It’s crucial to recognize that these figures are sourced directly from OpenAI and only reflect its own customer base, serving more as a measure of the company’s commercial health rather than an independent market study. Let’s dive into the details!

ChatGPT Enterprise Sees 800% Growth in a Year

The key figures from the report are quite revealing. The weekly message volume on ChatGPT Enterprise has increased eightfold from November 2024 to November 2025. The number of ChatGPT Enterprise seats has grown ninefold in the same period, now equipping 7 million workstations worldwide.

As for the API, the growth is even more remarkable. The average consumption of reasoning tokens per organization has skyrocketed by 320 times over one year. Over 9,000 organizations have now processed more than 10 billion tokens, and nearly 200 have surpassed the trillion-token mark. These volumes indicate a shift from experimentation to systematic integration into products and services.

“Enterprise AI is now entering this phase, as some of the world’s largest and most complex organizations begin to use AI as a central infrastructure,” writes OpenAI.

Saving 40 to 60 Minutes Daily According to OpenAI

OpenAI bases its findings on a survey conducted with 9,000 employees from nearly 100 companies to document productivity gains. According to their data, 75% of users report that AI improves the speed or quality of their work. ChatGPT Enterprise users claim to save between 40 and 60 minutes per active day using the tool.

The time savings vary by function. Professions in data science, engineering, and communications report saving 60 to 80 minutes daily. Accounting and finance roles appear to benefit most from each interaction with the AI.

Beyond time savings, 75% of surveyed workers state they can perform tasks they were unable to handle before. For instance, coding and data analysis are now extending beyond the strict confines of developers and analysts. Outside of engineering, IT, and research roles, the use of code-related messages has increased by an average of 36% over the last six months, estimates OpenAI.

France Among the Top 5 in Global Growth

Geographically, adoption is accelerating significantly outside the United States. Among the key markets, Australia shows the fastest growth with a 187% increase in the number of paying clients from November 2024 to November 2025, followed by Brazil (161%), the Netherlands (153%), and France (146%), all surpassing the global average increase of 143%.

In terms of message volume, the USA, Germany, and Japan remain the most active markets. The United Kingdom and Germany are among the largest markets for ChatGPT Enterprise outside the USA in terms of number of clients. In API usage, Japan is the leading market outside the USA in terms of corporate clients, with international growth exceeding 70% over the past six months.

A Growing Divide Between Leaders and Laggards

The report highlights an increasing gap between the most advanced users/companies and the median. The top 5% of AI users, termed “frontier users” by OpenAI, send six times more messages than the average user. In data analysis roles, they use the analytical tool 16 times more than the median. The difference is particularly stark in coding tasks, where frontier users send 17 times more coding messages than the median. For writing and communication, the ratio is 11, and 10 for analysis and calculations.

This disparity directly impacts performance. Users engaging the AI in about seven different types of tasks report saving five times more time than those using it for just four types. Among active monthly users, 19% have never used data analysis, 14% have never used reasoning, and 12% have never used research. These proportions drop to 3% and 1% among daily users.

Among businesses, frontier organizations generate about twice as many messages per seat as the median company, and seven times more messages to custom GPTs. These gaps suggest very different levels of organizational integration.

“Looking ahead, the next phase of enterprise AI will be shaped by improved performance on high-economic-value tasks, better understanding of organizational context, and a shift from simple result requests to models to delegating complex multi-step workflows,” predicts OpenAI.

Deeper Integration into Workflows

The use of custom GPTs and Projects has increased nineteenfold since the start of the year in terms of weekly active users. About 20% of all Enterprise messages now go through a custom GPT or a Project. These tools allow for configuring tailor-made assistants with specific instructions, knowledge, and customized actions to automate repetitive multi-step tasks.

Some organizations deploy these tools on a large scale. BBVA, for instance, regularly uses over 4,000 different GPTs, showcasing a culture where AI has become a daily, structured, and shared work tool across the enterprise.

Developers are quickly adopting Codex, too. In the last six weeks, the number of active weekly users of the coding assistant has doubled, and the volume of weekly messages has increased by about 50%. Teams are embracing the tool for code generation, refactoring, testing, and debugging, OpenAI details.

While these data points are impressive, they need to be contextualized. They reflect only the OpenAI ecosystem and are sourced from the company itself. Nonetheless, they are indicative of the accelerated adoption of AI in business, clearly moving from a phase of experimentation to integration into production workflows.