A Guide To Reading OpenAI’s EU Business Agentic AI Claim

🔍 Read the full analysis: A Guide To Reading OpenAI’s EU Business Agentic AI Claim on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get the latest gadgets 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 on October 14 that its coding agent 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. The figures offer a view of activity among OpenAI’s business customers, but they do not establish that agents handle most workplace AI tasks or create most business value.

OpenAI Economic Research’s report, Europe’s Agentic Opportunity: Moving from Breadth to Depth with AI Agents, defines agentic use as activity with Codex and ChatGPT Work, an agent mode launched in July 2026 that can work with files, browsers and company tools. The report says Codex accounted for 34% of output tokens and ChatGPT Work for 21% by September 30. The remaining 45% came from standard ChatGPT use in the covered accounts.

Among businesses with 200 or fewer users, OpenAI reports that weekly active users of Codex or ChatGPT Work rose 8.1 times between the week of May 1–7 and the week ending October 5. The company says every EU member state recorded at least fivefold growth; Belgium, Italy and Luxembourg had the largest reported multipliers. OpenAI does not give absolute user totals for this measure.

The report also finds wide differences in agentic output share across countries, from 85% in Lithuania and 81% in Cyprus to 22% in Slovenia and 26% in Austria. Germany’s share was 51%, ranking 17th among the 27 member states. These are figures for OpenAI’s accounts, not a census of AI use across European businesses.

What the Token Share Measures

The headline figure indicates where OpenAI-generated output is being produced inside its covered business accounts. It is relevant to companies tracking adoption of agent products and to policymakers weighing whether European firms are moving from occasional AI experiments toward more sustained use. But tokens measure text generated by a model, not completed work, time saved, accuracy or financial returns.

OpenAI’s Enterprise Signals page describes tokens as an imperfect measure of business value and says agentic tasks can require substantially more computation. Agents may read files, plan, generate material, check results and repeat steps; these processes can generate far more output than a short chat reply. That means a 55% token share should not be translated into a 55% share of interactions or jobs performed by agents.

The small-business growth rate also needs its time frame and product context. The comparison begins in early May, before ChatGPT Work launched in July. OpenAI’s report says a major part of the increase followed that launch. The rise indicates growing use of these tools within the measured customer group, but the report does not show how many businesses adopted them, how often they use them for valuable work, or whether the pattern extends to firms that do not buy OpenAI products.

How OpenAI Defines Agent Use

The October 14 materials include a Global Affairs post, the ten-page Economic Research report, and a global Enterprise Signals page. Together, they describe activity in ChatGPT Enterprise and Business accounts; they do not measure all European AI systems or the entire EU economy. The source data are therefore useful as a vendor-specific snapshot, rather than a representative survey of companies across Europe.

OpenAI compares the top 10% of firms by output tokens per active user with other firms. In September, it says EU frontier firms generated 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 group. The report estimates that EU companies account for about 10% of frontier firms worldwide, compared with about 65% in North America. These comparisons use OpenAI’s account activity and definitions, not independently audited measures of productivity or AI capability.

OpenAI frames its policy conclusion as moving from broader adoption to deeper use, including worker skills, confidence and access to advanced tools. The figures can inform that discussion, but they do not by themselves establish why EU firms trail the company’s global frontier measure or which policy changes would close the gap.

Limits of the EU Account Data

OpenAI does not publish country-level account or user counts alongside its country comparisons, making it hard to judge how representative each figure is or how much it may be shaped by a small number of large customers. The company itself notes that smaller countries can have relatively high agentic use within larger firms. The rankings should not be treated as a measure of national readiness.

The report also gives no absolute user counts for the 8.1-fold small-business increase and no independent measure of task completion, quality, productivity or return on investment. Its token figures cannot answer whether agent use has displaced chat, added new work, or produced better results. Nor does the supplied material establish how much growth would have occurred without ChatGPT Work’s launch.

OpenAI’s claims concern its own products and accounts. The source material does not provide an external audit, a comparison with other AI vendors, or evidence that the trends represent all EU businesses. Those limits do not make the account data meaningless; they define what the figures can support.

Evidence Needed Beyond Adoption

The immediate next step is to see whether OpenAI publishes follow-up data with absolute user counts, clearer country samples and longer time series. Such detail would help readers distinguish expansion from a small starting base and assess whether the country differences persist after the initial product-launch period.

For the broader question of economic impact, adoption metrics need to be paired with evidence about tasks completed, quality, time saved and costs. Until that information is available, the October report is best read as a snapshot of agent activity in OpenAI’s EU business accounts and a statement of the company’s policy priorities—not proof that agents now perform most work-related AI activity in Europe.

Key Questions

What does OpenAI mean by agents producing 55% of output?

It means Codex and ChatGPT Work generated 55% of output tokens in the covered EU ChatGPT Enterprise and Business accounts by September 30, 2026. It is not a measure of the share of users, tasks, or business value.

Which products count as agentic in the report?

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

Does the report show that most European companies use AI agents?

No. The figures cover OpenAI’s ChatGPT Enterprise and Business accounts in EU member states, not all European companies or all AI services.

How should readers interpret the 8.1-fold growth figure?

It is the reported increase in weekly active users of Codex or ChatGPT Work among businesses with 200 or fewer users, comparing the week of May 1–7 with the week ending October 5. OpenAI does not provide absolute user totals, and it says much of the growth followed ChatGPT Work’s launch.

What does the report say about Europe’s AI frontier?

Using output tokens per active user, OpenAI says EU frontier firms produced about four times the output of typical EU firms in September, while frontier firms outside the EU produced more than twice as much as the EU frontier group. This is a comparison of OpenAI account activity, not an independent productivity measure.

Source: ThorstenMeyerAI.com