OpenAI’s EU Business AI Claim: Agentic By Share, But Of What?

🔍 Read the full analysis: OpenAI’s EU Business AI Claim: Agentic By Share, But Of What? 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 said its AI agents generated 55% of output tokens in ChatGPT Enterprise and ChatGPT Business accounts across the 27 EU member states by September 30, according to a report published October 14. The figure combines Codex, its coding agent, and ChatGPT Work, but measures generated text rather than the share of business tasks, users or economic value handled by agents.

OpenAI Economic Research’s report, Europe’s Agentic Opportunity: Moving from Breadth to Depth with AI Agents, counts Codex as responsible for 34% of output tokens and ChatGPT Work for 21%. The remaining 45% came from standard ChatGPT. OpenAI defines “agentic” use in this report as activity from those two products.

The report also says weekly active users of Codex or ChatGPT Work among businesses with 200 or fewer users grew 8.1 times from the first week of May to the week ending October 5. Every EU member state recorded at least fivefold growth by this measure. OpenAI does not provide the absolute number of users behind that increase.

Reported agentic output shares differed considerably by country, from 85% in Lithuania and 81% in Cyprus to 22% in Slovenia and 26% in Austria. Germany was at 51%, ranking 17th among the 27 member states. OpenAI also reported that its EU frontier firms—its top 10% of firms by output tokens per active user—produced about four times as many tokens per user as typical EU firms. Firms in the global frontier outside the EU produced more than twice the per-user output of the EU frontier.

What the 55% Share Measures

The report offers a view of how OpenAI’s business customers use its products, and suggests that agent tools account for a substantial portion of generated text in those accounts. That may matter to companies deciding whether to expand access to coding and workplace agents, as well as policymakers weighing Europe’s ability to adopt advanced AI.

But output-token share is not a direct measure of adoption across workers or businesses. An agent can read files, plan, generate material and revise it, producing far more tokens than a short chat response. OpenAI itself calls tokens “an imperfect measure of business value” and notes that agentic tasks can require substantially more computation. The 55% result therefore describes where model output is being generated, not whether agents handle most work or create most value.

The report’s policy argument is that Europe should turn broader access into deeper use by building worker skills and confidence and making advanced tools more accessible and affordable. The data indicate a gap between the EU’s most intensive business users and the global frontier, but they do not establish that the gap is caused by any single policy, skill shortage or cost barrier.

A Product Launch Shapes the Trend

The two agent products have different histories. Codex was already in use in May; OpenAI launched ChatGPT Work in July 2026, adding an agent mode that can work with files, browsers and company tools. In OpenAI’s chart, Codex’s share rose from roughly a fifth of EU output tokens in May to 34% by September 30, while ChatGPT Work contributed 21 percentage points after its launch.

That timing matters when interpreting the reported growth. OpenAI says a major part of the increase in small-business weekly active users followed the introduction of ChatGPT Work. The 8.1-fold comparison starts before the product existed, so it combines adoption of an existing agent with uptake of a newly launched one. It shows a sharp rise in activity on OpenAI’s tools, but the report does not establish how much reflects lasting changes in business practice.

OpenAI’s report is based on its own Enterprise and Business accounts, not a survey of all EU companies. The company said EU firms make up about 10% of frontier firms worldwide, compared with about 65% in North America. Those figures, like the usage measures, are tied to OpenAI’s definitions and customer data.

Limits of OpenAI’s Customer Data

The report does not disclose country-level account or user counts, making it difficult to tell how representative the national comparisons are. OpenAI notes that smaller countries can have relatively high agent use within larger firms; a small number of large, agent-heavy customers could influence a country’s reported share.

It is also unclear how much agent use translates into completed tasks, time saved, improved output or revenue. The report gives no absolute user totals for the 8.1-fold small-business increase, and it does not separate the effects of ChatGPT Work’s launch from other changes in customer behavior. Nor do the figures show whether the activity is representative of businesses that do not use OpenAI products.

As a result, the country rankings should not be read as a ranking of national AI readiness. The measures capture different things: agentic output share and growth in weekly active users. The report itself gives no evidence that one measure predicts the other or that either one directly tracks business outcomes.

Evidence Needed Beyond Token Counts

OpenAI’s report calls for more skills, confidence and access to help European firms move from experimentation to more sustained AI use. The next test will be whether the company or independent researchers publish evidence connecting agent activity to specific work tasks and business outcomes, rather than usage volume alone.

Further reporting would also benefit from absolute user counts, clearer country-level sample sizes and comparisons that account for product launch dates. Until those details are available, the figures are best understood as a snapshot of activity within OpenAI’s EU business accounts through early October 2026—not a measure of agent adoption across Europe as a whole.

Key Questions

What does OpenAI mean by a 55% agentic share?

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 claim that agents handled 55% of tasks or users’ AI interactions.

How fast did agent use grow among smaller businesses?

OpenAI reported that weekly active users of Codex or ChatGPT Work in businesses with 200 or fewer users rose 8.1 times between the first week of May and the week ending October 5. The company did not publish the absolute user totals in the material provided.

Does the report show that agents create more business value than chat?

No. The report measures output tokens, not productivity, revenue or task completion. OpenAI describes tokens as an imperfect measure of business value, and agent tasks can generate more text and use more computation than chat replies.

Do the country figures rank EU countries by AI readiness?

No. They reflect usage in OpenAI’s own business accounts, and the report does not provide country-level account or user counts. The figures should not be treated as a representative ranking of all firms in each country.

Source: ThorstenMeyerAI.com