A Closer Look At OpenAI’s Agentic AI Claim For EU Businesses

🔍 Read the full analysis: A Closer Look At OpenAI’s Agentic AI Claim For EU Businesses on ThorstenMeyerAI.com

Age 18–24?Offer from Amazon

Prime made for students and young adults

• Fast, free delivery for dorm and study essentials
• Prime Video and Amazon Music included
• Member-only deals

Try Prime for Young Adults →Free trial for eligible 18–24 year olds

As an affiliate, we earn on qualifying purchases.Hide

OpenAI said Codex and ChatGPT Work generated 55% of output tokens in its ChatGPT Enterprise and Business accounts across the 27 EU member states by September 30, 2026. The figure, in a report published October 14, describes activity among OpenAI’s own business customers; it does not establish that agents handle most business AI tasks or generate most of the value from AI in Europe.

OpenAI Economic Research’s report, Europe’s Agentic Opportunity: Moving from breadth to depth with AI agents, defines agentic use as output from Codex, OpenAI’s coding agent, and ChatGPT Work, an agent mode introduced in July 2026 that can work with files, browsers and company tools. By September 30, Codex accounted for 34% of output tokens and ChatGPT Work for 21%; ordinary ChatGPT accounted for the remaining 45%.

For businesses with 200 or fewer users, OpenAI reports that weekly active users of Codex or ChatGPT Work rose 8.1 times between the first week of May and the week ending October 5. The report says every EU member state recorded at least fivefold growth. Belgium, Italy and Luxembourg had the highest reported growth, at 12.1 times, 10.8 times and 10 times, respectively. OpenAI did not provide absolute user totals in the supplied report summary.

Agentic output shares varied considerably by country: Lithuania was at 85% and Cyprus at 81%, while Slovenia was at 22% and Austria at 26%. Germany’s share was 51%, placing it 17th among the 27 member states. These are figures for OpenAI accounts, not national surveys of all businesses.

What the Token Share Shows

The figures offer a rare public view of how one major AI provider’s business customers are using its agent products. If the pattern holds across a wider set of firms, it could point to a shift from asking a model for a response toward giving it work that involves multiple steps and business tools. For companies, that shift can bring new questions about staff skills, oversight, access to company information and the reliability of automated work.

But output tokens are not a measure of business outcomes. OpenAI’s Enterprise Signals page describes tokens as an imperfect measure of value and says agentic tasks can require substantially more computation. An agent may read files, plan, produce material and check or revise its work, generating more text than a short chat exchange. A larger token share can therefore reflect the length and computational demands of agent runs, rather than a comparable share of tasks, users or economic benefit.

The report’s comparison of leading firms also points to a gap in usage intensity, not a direct measure of productivity. OpenAI defines “frontier firms” as the top 10% by output tokens per active user. It says EU frontier firms generated about four times as many tokens per user as typical EU firms, while frontier firms outside the EU generated more than twice as many as the EU frontier. That comparison raises questions about how deeply companies use the tools, but does not show what results those firms achieved.

How OpenAI Defines Agentic Use

The report covers ChatGPT Enterprise and ChatGPT Business accounts in the EU’s 27 member states. It is a provider-side account of product activity, rather than a survey of European firms as a whole. OpenAI published it through its Global Affairs team alongside its Economic Research report; its Enterprise Signals page supplies a broader global view.

Product timing matters when interpreting the growth numbers. Codex accounted for roughly a fifth of EU output tokens in May and 34% by September 30, according to the report’s chart. ChatGPT Work launched in July and contributed 21% by the end of September. The source says a major part of the small-business increase followed ChatGPT Work’s introduction. That means the reported rise reflects both adoption and the addition of a new product to the measure; it is not, by itself, evidence that business practices across Europe changed at the same pace.

OpenAI’s policy message is that Europe should move from broader AI adoption to deeper use by building workers’ skills and confidence and making advanced tools more accessible and affordable. The report does not establish that its account-level measures represent national readiness or the performance of European businesses overall.

Limits of the EU Account Data

The report does not show how many agentic tasks were completed, how many users relied on agents, or whether those tools improved productivity, quality or revenue. No task-level or business-outcome measure is supplied in the source material, and the 8.1-times growth figure is not accompanied by absolute user counts.

OpenAI also does not provide country-by-country account or user sample sizes in the supplied material. That makes it difficult to judge how representative the national figures are, especially for smaller markets where a limited number of large or agent-heavy customers could affect the reported share. The country rankings should not be treated as a ranking of national AI readiness.

The comparison between EU and non-EU frontier firms is also limited to output tokens per active user. The report’s definition does not establish that higher usage translates into better results. It remains unclear whether the patterns extend to businesses using other AI providers or to firms that do not have OpenAI’s enterprise and business accounts.

What OpenAI’s Figures Cannot Settle

OpenAI’s report provides a dated snapshot rather than a forecast: the main output-token figures end on September 30, 2026, while small-business activity is tracked through the week ending October 5. The source material does not announce a next measurement date or a planned follow-up release.

The most useful next evidence would include absolute user and account counts, clearer country-level sample sizes, and measures connecting agent use to completed work or business outcomes. Comparisons across providers and industries would also show whether the trend extends beyond OpenAI’s customer base. Until those data are available, the report supports a narrower conclusion: agents account for a majority of generated tokens in these OpenAI EU business accounts, while the wider economic impact remains unmeasured.

Key Questions

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

It means Codex and ChatGPT Work together generated 55% of the text tokens produced in the covered ChatGPT Enterprise and Business accounts in the EU by September 30, 2026. It does not mean they handled 55% of users’ tasks or produced 55% of business value.

Which OpenAI products count as agents 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.

How fast did agent use grow among small businesses?

OpenAI reports an 8.1-times increase in weekly active users of Codex or ChatGPT Work among businesses with 200 or fewer users, comparing the first week of May with the week ending October 5, 2026. The supplied material does not give the absolute user totals.

Does the report show that European businesses are becoming more productive?

No. It reports product usage and token volumes, not measured changes in productivity, revenue or work quality. OpenAI itself cautions that tokens are an imperfect measure of business value.

Do the country figures rank EU countries by AI readiness?

No. They describe activity in OpenAI’s own customer accounts, and the supplied material does not provide country-level sample sizes. The figures should not be read as a measure of all firms or national readiness.

Source: ThorstenMeyerAI.com