The Context Behind OpenAI’s Claim Of Agentic AI Use In EU Business

🔍 Read the full analysis: The Context Behind OpenAI’s Claim Of Agentic AI Use In EU Business on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on office and shipping supplies

• Business-only prices and quantity discounts
• Tax-exempt purchasing
• Multiple users, one account, clear invoices

Create a free Business account →Free to register

As an affiliate, we earn on qualifying purchases.Hide

OpenAI said agents generated 55% of output tokens in its ChatGPT Enterprise and Business accounts across the 27 European Union member states by Sept. 30, 2026, a measure of model output rather than the share of tasks or business activity handled by agents. In a report published Oct. 14, the company also reported rapid growth in weekly active agent users at smaller businesses, while noting that Europe’s highest-use companies still trail comparable firms elsewhere.

The figures come from OpenAI Economic Research’s report, Europe’s Agentic Opportunity: Moving from Breadth to Depth with AI Agents, published alongside a Global Affairs post titled The EU’s Next Chapter in AI Is Agentic. The report covers ChatGPT Enterprise and ChatGPT Business accounts in the EU. OpenAI defines agentic use here as use of Codex, its coding agent, and ChatGPT Work, an agent mode launched in July 2026 that can work with files, browsers and company tools.

By Sept. 30, Codex accounted for 34% of output tokens and ChatGPT Work for 21%, with ordinary ChatGPT accounting for the remaining 45%. OpenAI also said weekly active users of either agent at businesses with 200 or fewer users rose 8.1 times between the first week of May and the week ending Oct. 5. The report provides the growth multiple, but not the underlying user totals. It says every member state recorded at least fivefold growth during that period.

Use differed by country: OpenAI reported agentic output shares of 85% in Lithuania and 81% in Cyprus, compared with 22% in Slovenia and 26% in Austria. Germany’s reported share was 51%, ranking 17th among the 27 countries. The report also defines “frontier firms” as the top 10% by output tokens per active user. In September, EU frontier firms produced about four times the tokens per user of typical EU firms; frontier firms outside the EU produced more than twice as many as the EU frontier group.

What the Token Share Shows

The report offers a snapshot of how OpenAI’s business customers are using two products it classifies as agents. It suggests that agent features account for a substantial share of the company’s generated output in those accounts and that use among its smaller-business customers has grown quickly. That may matter to companies weighing how to adopt AI tools and to policymakers focused on workforce skills and access to advanced systems.

But token volume is not a direct measure of economic impact. OpenAI’s Enterprise Signals page calls tokens an imperfect measure of business value and says agentic tasks can require substantially more computation. An agent may read documents, plan, produce work, check it and revise it, generating more tokens than a short chat exchange. The 55% figure therefore cannot establish that agents handled most interactions, tasks or decisions, or that they delivered more value than other AI use.

The distinction matters when interpreting the headline. The findings apply to OpenAI’s EU business accounts, not every European company or AI provider. They describe activity recorded on one vendor’s products; the report does not measure whether that use improved productivity, cut costs or changed employment.

Product Launches Shape the Growth

The two agent products entered the market at different times during the report’s measurement period. Codex was already in use in May, when it accounted for roughly a fifth of EU output tokens in OpenAI’s chart. ChatGPT Work launched in July 2026 and accounted for 21% by Sept. 30. The source article’s reading of the chart is that ChatGPT Work contributed more than half of the roughly 35-percentage-point rise in agentic token share from May to September.

That timing also affects the small-business growth comparison. The 8.1-times measure begins with the week of May 1–7, before ChatGPT Work launched. OpenAI says a major part of the subsequent growth followed the launch. The result documents increased activity in its products, but without absolute user counts or a longer comparison period, it cannot show how broadly adoption has spread across the EU small-business sector.

OpenAI frames the policy implication as a need to turn broader adoption into deeper use, including through worker skills and confidence and more accessible and affordable advanced capabilities. CFO Sarah Friar said more than half of AI use among the company’s EU enterprise and business customers “now involves these tools.” That statement, like the report’s metrics, concerns OpenAI customers and its definition of agent tools.

Limits of the EU Account Data

OpenAI does not publish the absolute number of active users behind the 8.1-times growth figure, or account and user counts for each country. Without those details, it is difficult to judge how much small starting bases or a small number of large customers may affect the growth rates and country comparisons. The report itself notes that smaller countries can have relatively high agentic use within their larger firms.

The figures also do not reveal what share of user interactions or workplace tasks involved agents, how much work those agents completed without human intervention, or whether their use produced measurable business gains. The source material’s illustration that token shares could correspond to a much smaller share of interactions assumes agent runs produce 10 to 30 times as many output tokens as chat replies; that is an example assumption, not a result reported by OpenAI.

It is also unclear whether the reported growth will continue beyond the period covered, how agent use compares across vendors, and whether the observed country differences reflect adoption, customer mix or other factors. The available figures should not be treated as a ranking of national AI readiness.

More Evidence on Business Impact

The report’s measures cover activity through Sept. 30, 2026, with the small-business user-growth comparison ending the week of Oct. 5. OpenAI has not, in the material provided, set out a date for a follow-up report or supplied additional account-level totals. The next useful evidence would include absolute user counts, clearer breakdowns by business size and country, and measures that connect agent use to completed work or business outcomes.

For now, readers can compare the published data with later reporting from OpenAI and independent sources, while keeping the distinction between product activity and demonstrated impact. Whether fast growth persists, reaches firms beyond OpenAI’s customer base or produces measurable productivity changes remains open.

Key Questions

What does OpenAI mean by agentic use?

For this report, OpenAI counts use of Codex and ChatGPT Work as agentic. ChatGPT Work is an agent mode launched in July 2026 that can work with files, browsers and company tools.

Does 55% mean agents handled most EU business tasks?

No. The 55% figure is the share of output tokens generated by the two agent products in OpenAI’s EU business accounts. It does not measure the share of tasks, users, interactions or business value.

How did small-business agent use change?

OpenAI reported that weekly active users of Codex or ChatGPT Work at businesses with 200 or fewer users rose 8.1 times from the first week of May to the week ending Oct. 5, 2026. It did not publish the absolute user totals in the source material.

Can the country figures show which EU countries lead in AI readiness?

Not on their own. The report gives product-use figures but no country-level account or user counts, and the results may reflect differences in OpenAI’s customer mix. They should not be read as a definitive ranking of national AI readiness.

Source: ThorstenMeyerAI.com