OpenAI Says Agentic AI Leads In EU Business. First, Check What Was Counted

🔍 Read the full analysis: OpenAI Says Agentic AI Leads In EU Business. First, Check What Was Counted on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on monitors, keyboards and dev gear

• 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 Codex and ChatGPT Work generated 55% of output tokens in its ChatGPT Enterprise and ChatGPT Business accounts across the European Union by Sept. 30, 2026. The report also says weekly active use of those tools at businesses with 200 or fewer users rose 8.1 times from the first week of May, but the measures cover OpenAI customers and do not show that agents account for most AI tasks or business value across Europe.

OpenAI’s Global Affairs team published The EU’s next chapter in AI is agentic on Oct. 14 alongside a ten-page report from OpenAI Economic Research, Europe’s Agentic Opportunity: Moving from breadth to depth with AI agents. The analysis covers business accounts in the 27 EU member states. OpenAI defines agentic use here as output from Codex, its coding agent, and ChatGPT Work, an agent mode launched in July 2026 that can work with files, browsers and company tools.

As of Sept. 30, Codex accounted for 34% of output tokens and ChatGPT Work for 21%, while standard ChatGPT accounted for the remaining 45%. The report says weekly active users of Codex or ChatGPT Work at businesses with 200 or fewer users increased 8.1 times between the first week of May and the week ending Oct. 5. OpenAI reports at least fivefold growth in every member state, with Belgium at 12.1 times, Italy at 10.8 and Luxembourg at 10.0. It does not provide absolute user counts.

OpenAI also compares high-use customers. It defines “frontier firms” as the top 10% of firms by output tokens per active user. In September, 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. OpenAI says EU companies represent about 10% of frontier firms worldwide, compared with about 65% in North America.

What the Token Share Shows

The report offers a view into how OpenAI’s business customers are using its newer tools and where the company sees room for deeper adoption. But output tokens measure generated text, not the number of people using agents, the proportion of work completed with them, or the economic value they create. OpenAI’s Enterprise Signals page calls tokens an imperfect measure of business value and says agentic tasks can require substantially more computation.

That distinction matters because agent runs may read files, plan, generate code or documents, check results and repeat steps. Those activities can produce far more tokens than a short chat response. The report’s 55% figure therefore supports a narrower conclusion: agents account for more of the generated output in the covered accounts than standard ChatGPT does. It cannot, on its own, establish that most interactions or business activity are agentic.

The adoption figures also have a product-launch caveat. ChatGPT Work was not available in May, and OpenAI says a major part of the increase in small-business weekly active users followed its introduction. The reported growth indicates rising use from a May baseline, but the absence of absolute user counts makes it difficult to judge the scale of adoption or how broadly it extends beyond OpenAI’s customer base.

How OpenAI Defines Agentic Use

The headline figures combine two products: Codex and ChatGPT Work. According to the report’s figures, Codex’s share of EU business-account output tokens rose from roughly one-fifth in May to 34% by Sept. 30. ChatGPT Work contributed 21 percentage points after its July launch. That product timing means a substantial part of the increase occurred as OpenAI introduced a new tool, rather than measuring a change in use of a fixed set of products over the whole period.

Country results vary sharply: the report puts agentic output-token shares at 85% in Lithuania and 81% in Cyprus, compared with 22% in Slovenia and 26% in Austria. Germany is at 51%, 17th among the 27 member states. OpenAI provides no country-level account or user counts, and notes that some smaller countries have relatively high agentic use within their larger firms. These are results for OpenAI’s customer mix, not a ranking of national AI readiness.

The report’s stated policy argument is that Europe should move from broad AI adoption to deeper use by building workers’ skills and confidence and making advanced capabilities more accessible and affordable. That is OpenAI’s interpretation of its customer data; the report does not establish that its account-level measures represent all European companies.

Limits in the EU Customer Data

The report does not give absolute account or user counts for its small-business growth measure or its country comparisons. It is therefore not possible from the published figures to determine how many businesses drove the increases, how representative the accounts are of each country’s economy, or how much the starting sample size affected the growth multiples.

OpenAI’s data also do not establish how much work agents completed or whether their use improved productivity, costs or other business outcomes. The report’s 55% token share cannot answer those questions. The source material also describes five caveats but cuts off before completing its final point, so no additional conclusion from that section can be verified here.

Measuring Adoption Beyond Tokens

The published reports provide a snapshot through Sept. 30 for output-token shares and through the week ending Oct. 5 for small-business activity. No further release date or independent validation is specified in the source material. A fuller assessment would need later data and measures that clarify the number and type of businesses represented, the frequency and purpose of agent use, and the work or outcomes those tools support.

For now, the firmest reading is limited to OpenAI’s own accounts: agentic tools produced more than half of the output tokens in the EU business accounts covered by the report, and weekly active use rose from the stated May baseline. Whether that pattern reflects wider adoption across European business remains unestablished.

Key Questions

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

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

Which tools did OpenAI count as agentic?

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.

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

OpenAI reports 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. It gives no absolute user count and says a major part of the growth followed ChatGPT Work’s launch.

Do the country figures measure national AI adoption?

No. They describe OpenAI’s business accounts in each member state. The report does not provide country-level sample sizes, so the figures should not be treated as a ranking of all businesses or national AI readiness.

Source: ThorstenMeyerAI.com