OpenAI Says EU Businesses Have Shifted To Agentic AI—Look At The Numbers

🔍 Read the full analysis: OpenAI Says EU Businesses Have Shifted To Agentic AI—Look At The Numbers on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get hardware and tech essentials delivered free with Prime

• Fast, free delivery on millions of items
• Prime Video, Amazon Music and more included
• Member-only deals all year

Start your free Prime trial →Free trial for eligible customers · Cancel anytime

As an affiliate, we earn on qualifying purchases.Hide

OpenAI says AI agents generated 55% of output tokens in its ChatGPT Enterprise and ChatGPT Business accounts across the European Union by Sept. 30, a sign that use of its agent products is growing among business customers. The company’s Oct. 14 report also says weekly active users of those tools at businesses with 200 or fewer users rose 8.1× from the first week of May; the figures describe activity on OpenAI’s products, not adoption across all EU companies.

OpenAI’s Economic Research report, Europe’s Agentic Opportunity: Moving from Breadth to Depth with AI Agents, covers business accounts in the EU’s 27 member states. It defines “agentic” use as activity involving Codex, OpenAI’s coding agent, and ChatGPT Work, an agent mode that launched in July 2026 and can work with files, browsers and company tools.

By Sept. 30, Codex accounted for 34% of output tokens in the covered accounts and ChatGPT Work for 21%, putting their combined share at 55%. OpenAI says ordinary ChatGPT use accounted for the remaining 45%. The company’s global Enterprise Signals material cautions that token counts are an imperfect measure of business value and that agent tasks can use substantially more computation than simpler interactions.

For businesses with 200 or fewer users, OpenAI reports that weekly active users of Codex or ChatGPT Work increased 8.1 times between the first week of May and the week ending Oct. 5. It says every EU member state recorded at least fivefold growth. Belgium had the largest reported increase, at 12.1 times, followed by Italy at 10.8 times and Luxembourg at 10.0 times. OpenAI does not provide absolute user counts alongside those growth rates.

What the Token Share Can Show

The figures offer a view of how customers of one major AI provider are using its newer tools. If agents perform more work across code, documents and connected tools, that could affect how companies plan software development and other office processes. OpenAI’s numbers indicate that agent products account for a large share of the text generated in its EU business accounts, and that activity increased after ChatGPT Work became available.

But 55% of output tokens is not 55% of tasks, users or economic value. An agent may read and revise files, plan steps and check its own work, generating many more tokens than a brief chat response. That means the output-token share cannot establish that most business interactions are agentic or that agents deliver most of the value. OpenAI itself describes tokens as an imperfect measure of business value.

The distinction matters for businesses and policymakers deciding whether the data reflects a broad change in workplace practice. The report describes activity within OpenAI’s paid business accounts; it does not measure the full EU economy or compare the outcomes of agent use with other ways of working. The figures are best read as product-usage data, not a survey of European companies as a whole.

A Product Launch Shapes the Trend

The reported growth period begins in early May, before ChatGPT Work launched in July. OpenAI’s report says a major part of small-business growth followed the introduction of that agent mode. The product launch is relevant when interpreting the 8.1× figure: the comparison spans a period when the set of available agent tools changed, and OpenAI supplies no absolute user counts to show how many people the multiple represents.

OpenAI also reports wide differences among EU countries. The agentic share of output tokens ranges from 85% in Lithuania and 81% in Cyprus to 22% in Slovenia and 26% in Austria. Germany is reported at 51%, ranking 17th among the 27 member states. These are shares in OpenAI’s covered accounts, not national rates of AI adoption.

The report defines frontier firms as the top 10% of companies by output tokens per active user. In September, OpenAI says EU frontier firms produced about four times as many tokens per user as typical EU firms, while frontier firms outside the EU produced more than twice as many as the EU frontier group. It estimates EU companies make up about 10% of frontier firms worldwide, compared with about 65% in North America. Those comparisons use OpenAI’s own customer data and its stated definition of frontier firms.

Limits of OpenAI’s Customer Data

The report does not disclose the number of accounts or users behind each country result, or the absolute number of small-business users represented by the 8.1× increase. Without those details, it is difficult to tell how much country rankings reflect broad usage patterns and how much they reflect the composition of OpenAI’s customer base. The results also do not establish whether users completed more work, saved time or produced better outcomes.

OpenAI’s figures cover its own Enterprise and Business accounts, rather than a representative survey of companies across Europe. The period includes the July launch of ChatGPT Work, and the report attributes a major part of small-business growth to that launch. The data therefore does not, on its own, show that agent use has become a structural feature of the EU small-business economy. The source material also does not provide a comparable measure of agent use across other AI providers.

Evidence Needed Beyond Usage

The report sets out OpenAI’s view that Europe should turn wider adoption into deeper use by supporting worker skills and confidence and making advanced tools more accessible and affordable. It does not specify a timetable or provide a measured forecast for how quickly firms will expand agent use.

The next useful evidence would include absolute user and account counts, clearer information about country samples, and measures of tasks completed and business outcomes over time. Further reporting should also test whether growth continues beyond the period following ChatGPT Work’s launch and whether similar patterns appear across providers. Until then, OpenAI’s data shows a sharp rise in activity in its own business accounts, while the scale and value of agent adoption across European businesses remain unsettled.

Key Questions

What does OpenAI mean by “55% of output tokens”?

It means Codex and ChatGPT Work together generated 55% of the text output tokens in the EU ChatGPT Enterprise and Business accounts covered by the report, through Sept. 30, 2026. It is not a measure of the percentage of users, tasks or business value.

How much did small-business agent use grow?

OpenAI reports that weekly active users of Codex or ChatGPT Work at businesses with 200 or fewer users rose 8.1× from the first week of May to the week ending Oct. 5. The report does not give the absolute user count behind that multiple.

Which agent tools are included in the report?

The report counts Codex, OpenAI’s coding agent, and ChatGPT Work, an agent mode launched in July 2026 that can work with files, browsers and company tools.

Do the country figures show which EU nations use AI agents most?

They show agentic output-token shares and reported growth among OpenAI’s covered business accounts in each country. OpenAI does not provide country-level sample sizes, so the figures should not be treated as rankings of AI adoption across entire national economies.

Source: ThorstenMeyerAI.com