OpenAI’s EU Business AI Claim: Start By Asking “Most Of What?”

🔍 Read the full analysis: OpenAI’s EU Business AI Claim: Start By Asking “Most Of What?” 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 its AI agents generated 55% of output tokens in its ChatGPT Enterprise and ChatGPT Business accounts across the European Union by Sept. 30, according to a report published Oct. 14. The figure tracks text generated by Codex and ChatGPT Work, not the share of customers’ tasks or the economic value they gained, so it offers a narrower measure of adoption than the headline may suggest.

The report, Europe’s Agentic Opportunity: Moving from breadth to depth with AI agents, was published by OpenAI Economic Research alongside a post from the company’s Global Affairs team. It covers accounts in the EU’s 27 member states. OpenAI defines “agentic” use in this analysis as work done by 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%, while standard ChatGPT accounted for the remaining 45%. OpenAI CFO Sarah Friar said that more than half of AI use among the company’s EU enterprise and business customers “now involves these tools.” In this report, however, the underlying headline statistic is specifically about output tokens.

OpenAI also reported an 8.1-fold rise in weekly active users of Codex or ChatGPT Work among businesses with 200 or fewer users. The comparison runs from the first week of May to the week ending Oct. 5. OpenAI said every EU member state recorded at least fivefold growth; Belgium led at 12.1-fold, followed by Italy at 10.8-fold and Luxembourg at 10-fold. The report does not give absolute user counts for that measure.

What the Token Share Can Show

The figures show that agent products account for a substantial share of the text generated inside OpenAI’s EU business accounts. They may help businesses and policymakers track how quickly customers are trying newer tools. But output volume is not a direct measure of adoption across the economy: a relatively small number of long-running agent sessions could generate many more tokens than brief chat exchanges.

OpenAI’s own Enterprise Signals material calls tokens an imperfect measure of business value and notes that agentic tasks can require substantially more computation. An agent may read files, plan, draft code or documents, check results and repeat steps. Those activities can produce far more output than a short chat response, without establishing that agents are handling most tasks or delivering greater value.

The distinction matters when comparing countries or drawing policy conclusions. The report’s central recommendation is to move from broad adoption toward deeper use through worker skills, confidence and access to advanced capabilities. That is a policy argument from OpenAI, not a finding that the token data alone proves a productivity gain or a national readiness ranking.

A New Agent Mode Shapes the Trend

The reported rise spans a period when OpenAI introduced a new product. ChatGPT Work launched in July 2026, after the small-business comparison’s May starting week. In the report’s chart, Codex made up roughly one-fifth of EU output tokens in May and 34% by Sept. 30; ChatGPT Work contributed 21 percentage points by the end of that period.

That timing complicates the interpretation of the growth figures. OpenAI says a major part of small-business growth followed ChatGPT Work’s introduction. The 8.1-fold increase therefore records expanding use of the company’s agent products from a pre-launch starting point, but does not by itself show that the same pace will continue or that it reflects a lasting change across European businesses.

OpenAI also compared firms by output tokens per active user. It classified the top 10% as “frontier firms.” 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 report says EU companies account for about 10% of frontier firms worldwide, compared with about 65% in North America. These are measures based on OpenAI’s customers and its token-based definition, not a census of all companies.

What the Data Leaves Out

The report does not establish what share of all AI interactions, users or business tasks in Europe involve agents. It also provides no direct measure of productivity, cost savings or business value. OpenAI’s customer accounts are not the same as a representative sample of European companies, and the supplied material does not give country-level account or user counts.

Those missing sample sizes make the national comparisons difficult to interpret. OpenAI reported agentic output shares ranging from 85% in Lithuania to 22% in Slovenia, with Germany at 51%. The company also notes that smaller countries such as Cyprus, Estonia and Lithuania have relatively high agent use within larger firms. Without account counts and customer-mix details, the figures cannot show whether national differences reflect broad business trends or a small number of heavy users.

The timing of the product launch is another limit. ChatGPT Work was introduced during the period measured, and OpenAI says it contributed substantially to growth. The figures do not show how much use would have risen without that launch, whether users will keep using the tools, or how many people and companies the user-growth multiple represents in absolute terms.

Evidence Needed Beyond OpenAI Accounts

The report points to a continuing debate about whether Europe can turn initial AI adoption into more regular, effective workplace use. OpenAI’s stated priorities are workforce skills and confidence, along with making advanced capabilities more accessible and affordable. The company’s data can track activity in its own products, but independent evidence would be needed to assess uptake across European businesses and its effect on output, costs or jobs.

Readers should watch for later reporting that provides absolute user counts, clearer sampling details and comparisons over time after ChatGPT Work’s launch period. Until then, the strongest supported conclusion is limited: agents generated more than half of the output tokens in these OpenAI business accounts by Sept. 30, while the wider implications for European business adoption and productivity remain unproven by these figures.

Key Questions

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

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

Which tools counted as agentic in the report?

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.

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-fold from the first week of May to the week ending Oct. 5. The report does not provide absolute user counts, and the comparison period includes ChatGPT Work’s launch.

Does the report prove that European businesses are becoming more productive?

No. It reports activity in OpenAI’s products, primarily through token counts and active-user growth. It does not directly measure productivity, cost savings or business value, and it does not represent every European company.

Source: ThorstenMeyerAI.com