OpenAI Says Agentic AI Makes Up Most EU Business AI—Check The Sample

🔍 Read the full analysis: OpenAI Says Agentic AI Makes Up Most EU Business AI—Check The Sample on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on office and shipping supplies

• 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 says its Codex and ChatGPT Work agents produced 55% of output tokens in its ChatGPT Enterprise and Business accounts across the European Union by Sept. 30, a finding that points to growing agent use among the company’s business customers but does not show that agents handle most European business AI tasks. The figures come from a report published Oct. 14 by OpenAI Economic Research and cover OpenAI accounts in the 27 EU member states, not a representative survey of all European firms.

OpenAI’s report, “Europe’s Agentic Opportunity: Moving from breadth to depth with AI agents,” defines agentic activity as use of Codex, its coding agent, and ChatGPT Work, an agent mode that launched in July 2026 and can work with files, browsers and company tools. By Sept. 30, Codex accounted for 34% of output tokens and ChatGPT Work for 21%. The remaining 45% came from standard ChatGPT use, according to the report.

The company also reports a sharp rise among smaller businesses. At businesses with 200 or fewer users, weekly active users of Codex or ChatGPT Work increased 8.1 times between the week of May 1–7 and the week ending Oct. 5. OpenAI says every EU member state recorded at least fivefold growth over that period; Belgium, Italy and Luxembourg had the largest reported increases, at 12.1, 10.8 and 10.0 times, respectively. The report does not provide absolute user counts for this measure.

Agentic output varied across countries in OpenAI’s data. Lithuania had the highest reported share, at 85%, followed by Cyprus at 81%; Slovenia was at 22% and Austria at 26%. Germany’s share was 51%, ranking 17th among the 27 member states. These are proportions of output tokens within OpenAI’s covered accounts, not national adoption rates.

The report also compares high-use companies. OpenAI 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 estimates EU companies make up about 10% of frontier firms worldwide, compared with about 65% in North America.

What the 55% Token Share Shows

The report offers a window into how businesses using OpenAI’s paid products are adopting tools that can take multi-step actions, rather than only respond to prompts. The figures suggest that agent workflows account for a substantial share of generated text in those accounts, and that use of the company’s agent products expanded rapidly after ChatGPT Work launched.

But output tokens are not a direct measure of work completed or value created. An agent may read files, plan, generate code or documents, check results and retry, producing much more text than a short chat exchange. OpenAI’s Enterprise Signals page describes tokens as an imperfect measure of business value and says agentic tasks can require substantially more computation. The 55% figure therefore indicates where model output is being generated; it does not establish that agents perform 55% of tasks, serve 55% of users or deliver 55% of business value.

The distinction matters for employers and policymakers weighing whether agent adoption is broad enough to change productivity. The report’s evidence concerns one provider’s customers and products. It can inform discussion about usage patterns, but it cannot by itself establish how widely agents are used across the EU economy or whether they improve business outcomes.

Popular right now

Bose QuietComfort True Wireless Earbuds with Active Noise Cancellation, Bluetooth 5.3, USB-C Charging, IPX4, Black

AmazonView Latest Price

Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation

AmazonView Latest Price

Vitamix 5200 Blender with 64 oz Container, Black

AmazonView Latest Price

As an affiliate, we earn on qualifying purchases.

How OpenAI Defines Agent Use

OpenAI published two related materials on Oct. 14: its Global Affairs post, “The EU’s next chapter in AI is agentic,” and the ten-page Economic Research report. Its global Enterprise Signals page provides additional comparisons. The report’s scope is ChatGPT Enterprise and ChatGPT Business accounts in the EU’s 27 member states; it is not described as a survey of all companies or AI vendors.

The timing of product releases is relevant to the growth figures. OpenAI’s report says Codex accounted for roughly a fifth of EU output tokens in May and 34% by Sept. 30. ChatGPT Work, launched in July, contributed 21 percentage points by the end of September. OpenAI also says a major part of the small-business user growth followed that launch. The reported rise thus includes activity around a newly introduced product, rather than measuring a long-established trend alone.

OpenAI’s stated policy message is to turn wider AI adoption into deeper use by helping businesses move beyond experiments, build workers’ skills and confidence, and make advanced capabilities more accessible and affordable. That is the company’s interpretation and policy argument; the usage data does not independently test whether those measures would raise productivity.

Limits in the EU Account Data

OpenAI does not publish account or user sample sizes for each country in the supplied material, making it difficult to judge how representative country-level shares are. The company itself says smaller countries such as Cyprus, Estonia and Lithuania have relatively high agentic use within their larger firms. A small number of large or agent-heavy customers could therefore influence a national figure. The rankings should not be read as a definitive league table of national AI readiness.

The report also does not give absolute counts behind the 8.1-fold small-business growth, so readers cannot tell from that multiplier alone how many businesses or users it represents. The starting point is the first week of May, before ChatGPT Work launched, and OpenAI says the launch contributed substantially to the increase. The data does not separate product-launch effects from longer-term changes in adoption.

Nor does the report establish how many tasks agents complete, whether their use saves time or money, or whether the token-intensive activity generates better results. No independent, economy-wide comparison is supplied in the source material. The findings are best understood as a snapshot of OpenAI’s business accounts through early October, not a census of EU business AI.

Evidence Needed Beyond Usage

The next relevant evidence would include updated usage figures over a longer period, published absolute user counts and clearer country-level sample sizes. Comparisons across providers and independent surveys could help establish whether the patterns in OpenAI accounts reflect wider business adoption in Europe.

To evaluate economic impact, future research would also need measures beyond tokens and active users, such as tasks completed, time saved, quality, costs and the effect on workers. OpenAI’s report calls for skills development and broader access to advanced tools, but the supplied material does not set out a timetable for policy measures or a date for a follow-up report. For now, the figures document reported activity in OpenAI’s EU business products; broader conclusions remain open.

Key Questions

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

OpenAI says Codex and ChatGPT Work generated 55% of output tokens in its EU ChatGPT Enterprise and Business accounts by Sept. 30, 2026. The figure combines 34% from Codex and 21% from ChatGPT Work.

Does the report show that agents perform most AI tasks in Europe?

No. The 55% figure measures output tokens in OpenAI accounts, not the share of users, tasks or business value. Agent runs can generate more text than ordinary chat exchanges.

What does the 8.1-fold growth figure measure?

It measures growth in weekly active Codex or ChatGPT Work users at businesses with 200 or fewer users, from the week of May 1–7 to the week ending Oct. 5, 2026. OpenAI does not provide the absolute user count in the supplied report material.

Are the country figures national adoption rates?

No. They describe the agentic share of output tokens among OpenAI’s covered accounts in each country. OpenAI does not provide country-level sample sizes in the supplied material, and the figures do not represent all firms in those countries.

Does the report show that agents improve productivity?

No. The report describes usage patterns and OpenAI’s policy recommendations, but the figures cited do not measure productivity, cost savings or business value.

Source: ThorstenMeyerAI.com