A Denominator Check For OpenAI’s EU Business Agentic AI Figure

🔍 Read the full analysis: A Denominator Check For OpenAI’s EU Business Agentic AI Figure on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get smart everyday buys 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 reported that its agent products generated 55% of output tokens in its ChatGPT Enterprise and Business accounts across the 27 European Union member states by September 30, 2026. The figure comes from the company’s own customer data and measures text generated by its systems—not the share of European businesses, tasks or economic value accounted for by agents.

In a report published October 14, OpenAI Economic Research said its measure of “agentic” use includes Codex, its coding agent, and ChatGPT Work, an agent mode launched in July that can work with files, browsers and company tools. Codex accounted for 34% of output tokens in the covered accounts by September 30, while ChatGPT Work accounted for 21%. Standard ChatGPT use made up the remaining 45%.

OpenAI also reported rapid growth among smaller business accounts. For businesses with 200 or fewer users, weekly active users of Codex or ChatGPT Work increased 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 largest reported increases, at 12.1, 10.8 and 10 times, respectively. OpenAI did not provide absolute user counts for this measure.

The report found substantial differences between countries in the share of account output attributed to agents: Lithuania was at 85% and Cyprus at 81%, compared with 22% in Slovenia and 26% in Austria. Germany’s reported share was 51%, placing it 17th among the 27 member states. These are readings from OpenAI’s customer base, not estimates of all AI use in each country.

What the Token Share Captures

The figures offer a view of how customers are using OpenAI’s products, but output tokens are not a direct measure of work completed. An agent may read documents, plan steps, generate code or text, check its output and revise it. Those actions can produce substantially more tokens than a short chat exchange, even if they represent only one task or interaction.

OpenAI’s Enterprise Signals material describes tokens as an imperfect measure of business value and says agentic tasks can require substantially more computation. That means a majority share of tokens cannot establish that agents perform most workplace AI tasks, serve most users, or create most of the value. It does show that agents account for a large share of the text generated in these particular OpenAI accounts.

The distinction matters for companies weighing adoption and for readers interpreting claims about Europe’s competitiveness. The report presents an adoption signal and a comparison within OpenAI’s ecosystem. It does not independently establish productivity gains, cost savings or the readiness of European firms as a whole.

A Product Launch Shapes the Baseline

The reported growth period begins in early May, before ChatGPT Work launched in July 2026. OpenAI’s chart shows Codex rising from roughly a fifth of EU output tokens in May to 34% by September 30, while Work added 21 percentage points after its launch. The source article’s reading of the chart is that Work accounts for more than half of the roughly 35-point rise in agentic token share. OpenAI also says a major part of small-business growth followed Work’s introduction.

That timing does not make the adoption figures invalid, but it affects their interpretation: the comparison includes the rollout of a new product, not just gradual changes in established behavior. An increase of 8.1 times from the May baseline signals growing use among covered accounts, while the lack of absolute user counts makes it difficult to judge the scale of that increase or how representative it is of small businesses across Europe.

OpenAI’s report also compares high-usage companies. It 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, while frontier firms outside the EU generated more than twice as many as the EU frontier. OpenAI says EU firms make up about 10% of frontier firms worldwide, compared with about 65% in North America. These comparisons use the report’s token-based definitions and are not direct productivity rankings.

What the Data Cannot Establish

The report does not provide country-level account or user counts, making it difficult to tell how much a small number of large customers may influence national results. OpenAI itself notes that smaller countries including Cyprus, Estonia and Lithuania have relatively high agent use within their larger firms. The country figures should not be treated as a ranking of national AI readiness.

Nor does the report disclose absolute weekly active-user numbers behind the 8.1-times small-business increase. Without those counts, the starting scale and the number of businesses represented remain unclear. The token data also cannot show how many distinct tasks were agentic, whether those tasks were completed successfully, or whether they improved business outcomes.

The source material further notes that country rankings by agentic share and by small-business growth do not closely align. That weak relationship suggests the measures may reflect different customer mixes or stages of adoption, rather than a single, clear national trend. OpenAI’s published figures cover its Enterprise and Business accounts, not all EU firms or their use of competing AI services.

What to Watch in Future Data

The next useful test will be whether OpenAI publishes updated figures that include absolute user and account counts, clearer information about country sample sizes, and results over a longer period after ChatGPT Work’s launch. Such data could help distinguish continued growth from the initial effect of introducing the product.

Readers will also need measures beyond token volume to assess the business impact OpenAI’s report discusses: for example, task completion, time saved, costs and quality, with enough detail to compare results across firms. The report sets out a policy case for building worker skills and making advanced tools more accessible and affordable, but it does not provide economy-wide evidence that those steps have already produced gains.

Until further data are available, the confirmed takeaway remains narrow: agents generated more output tokens than standard ChatGPT in the covered OpenAI EU business accounts by September 30. Whether that usage represents a broader shift in European workplaces, and what value it creates, remains open.

Key Questions

What does OpenAI’s 55% figure measure?

It is the share of output tokens generated by Codex and ChatGPT Work in ChatGPT Enterprise and Business accounts across the EU by September 30, 2026. It is not the share of users, tasks or business value.

Which products does OpenAI count as agentic?

The report counts Codex, OpenAI’s coding agent, and ChatGPT Work, an agent mode launched in July 2026. Standard ChatGPT accounts for the remaining 45% of reported output tokens.

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

OpenAI reports an 8.1-times increase in weekly active users of Codex or ChatGPT Work among its business accounts with 200 or fewer users. The comparison runs from the first week of May to the week ending October 5. The company did not publish absolute user counts, and the figure does not cover every small business in the EU.

Does the report show that agents create most business value?

No. Token volume measures generated text, and agent runs can use more tokens than ordinary chat. OpenAI describes tokens as an imperfect measure of business value; the report does not establish the share of work completed or value created by agents.

Source: ThorstenMeyerAI.com