Is Agentic AI Really Most Of EU Business AI? OpenAI’s Denominator Matters

🔍 Read the full analysis: Is Agentic AI Really Most Of EU Business AI? OpenAI’s Denominator Matters on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on tech for your team

• 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 reported that its AI agents generated 55% of output tokens in ChatGPT Enterprise and Business accounts across the 27 EU member states by Sept. 30, 2026. The figure highlights rising use of tools that can work across files, browsers and company software, but it measures generated text—not the share of European businesses, tasks or economic value involving agents.

In its report, Europe’s Agentic Opportunity: Moving from Breadth to Depth with AI Agents, OpenAI Economic Research counted output from Codex, the company’s coding agent, and ChatGPT Work, an agent mode launched in July. Codex accounted for 34% of output tokens and ChatGPT Work for 21% by Sept. 30. Standard ChatGPT accounted for the remaining 45%.

OpenAI also said 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. The company said every EU member state recorded at least fivefold growth. It did not provide absolute user counts, so the rate does not show how many small-business employees adopted the tools.

The report found a wide spread in the agentic share of output across countries: Lithuania had 85% and Cyprus 81%, while Slovenia had 22% and Austria 26%. Germany recorded 51%, placing it 17th among the 27 member states. OpenAI did not publish country-level account or user sample sizes alongside those figures.

OpenAI’s comparison of high-output users also pointed to a gap between Europe and the global frontier. The report defines “frontier firms” as the top 10% by output tokens per active user. In September, EU frontier firms generated about four times as many tokens per user as typical EU firms; frontier firms outside the EU generated more than twice as many as the EU frontier. The EU accounted for about 10% of frontier firms worldwide, compared with about 65% in North America.

What the 55% Figure Measures

The headline number offers a view of how OpenAI’s business customers use its products, but it is not a direct measure of how work is changing across Europe. Output tokens measure generated text. An agent that reads files, plans a task, writes code and checks its work may generate much more text than a single chat response. OpenAI’s own Enterprise Signals page calls tokens an imperfect measure of business value and says agentic tasks can require substantially more computation.

That difference matters when interpreting both the 55% share and the 8.1-times growth figure. The token share cannot establish that agents handle most AI interactions, that most companies use them, or that they deliver more value than chat tools. The growth rate is measured from early May, before ChatGPT Work launched, and OpenAI says a major part of the increase followed that launch. Without absolute user counts or a comparison against other AI providers, the report describes activity in OpenAI’s customer base rather than the EU business economy as a whole.

For business leaders and policymakers, the report still points to a practical question: whether early use expands into sustained, productive use. OpenAI’s policy argument is that Europe should move from broad adoption toward deeper use by improving worker skills and confidence and making advanced tools more accessible and affordable. Those are recommendations from the company, not outcomes demonstrated by the usage figures.

How OpenAI Defines Agentic Use

The report was published on Oct. 14, 2026, by OpenAI’s Global Affairs team alongside research from OpenAI Economic Research. Its scope is limited to ChatGPT Enterprise and ChatGPT Business accounts in the 27 EU member states. It does not describe all European companies, consumer use, or activity on other AI platforms.

OpenAI groups Codex and ChatGPT Work under “agentic” use. ChatGPT Work launched in July 2026 and can operate with files, browsers and company tools; Codex is the company’s coding agent. That timing is relevant to the reported rise: Codex’s share of output tokens increased from roughly one-fifth in May to 34% by Sept. 30, while ChatGPT Work contributed 21 percentage points after its launch, according to the report’s chart.

The company’s stated policy conclusion is to turn wider adoption into deeper use. Its CFO, Sarah Friar, said that more than half of AI use among OpenAI’s EU enterprise and business customers “now involves these tools.” That statement describes OpenAI’s customers and its definition of agentic tools; it should not be read as a finding about all EU businesses.

Limits of the Customer Data

OpenAI did not disclose absolute account or user counts for the small-business growth calculation or sample sizes for the country comparisons. That makes it difficult to judge how representative the country figures are, particularly in smaller markets where a limited number of large or highly agentic customers could influence the share.

The report also does not establish how many tasks were handled by agents, how much time or money businesses saved, or whether the tools improved work quality. Its token measure can rise when agent runs generate more text per task, even without a comparable rise in the number of tasks. The 8.1-times figure starts in the week of May 1–7, before ChatGPT Work launched, and the source gives no absolute baseline user count.

These data describe OpenAI products and customers. The report does not provide a cross-provider comparison or a representative survey of all EU firms. It therefore cannot establish that agents now make up most business AI use across Europe, or that the frontier-firm gap reflects productivity rather than differences in customer mix, usage patterns or product access.

Evidence Needed Beyond Usage

The report does not announce a follow-up release date or a next measurement milestone. The next evidence needed to test its broader implications would include absolute adoption figures, consistent measures of tasks and outcomes, and data that can be compared across countries and providers. Those measures would help separate product-driven usage growth from wider changes in business practice.

For now, the figures provide a dated snapshot: token output through Sept. 30, frontier-firm comparisons for September, and small-business weekly active use through Oct. 5. OpenAI’s interpretation is that Europe has room to deepen adoption. Whether that leads to measurable business gains, and whether the reported gap closes, remains unshown by the report.

Key Questions

What does OpenAI mean by 55% of EU business AI being agentic?

It means Codex and ChatGPT Work generated 55% of output tokens in the ChatGPT Enterprise and Business accounts covered by OpenAI’s EU analysis by Sept. 30, 2026. It does not mean 55% of users, tasks, companies or business value involved agents.

Which tools did OpenAI count as agents?

OpenAI counted Codex, its coding agent, and ChatGPT Work, an agent mode launched in July 2026 that can work with files, browsers and company tools.

Did small-business agent use grow 8.1 times across Europe?

OpenAI reported an 8.1-times rise in weekly active users of Codex or ChatGPT Work at businesses with 200 or fewer users, comparing the first week of May with the week ending Oct. 5. It did not give absolute user counts, and the period includes ChatGPT Work’s launch.

Does the report show that European businesses lag the global AI frontier?

It reports that EU frontier firms generated about four times as many tokens per active user as typical EU firms, while frontier firms outside the EU generated more than twice as many as the EU frontier. This is a comparison of token output per user, not direct evidence of a productivity or business-value gap.

Can the country rankings show which EU nations are most ready for AI agents?

Not on their own. OpenAI reported country-level agentic output shares and growth rates but did not provide country sample sizes. The figures describe its customer base and should not be treated as a definitive ranking of national AI readiness.

Source: ThorstenMeyerAI.com