Before Drawing Conclusions From OpenAI’s EU Business AI Claim, Check The Base

🔍 Read the full analysis: Before Drawing Conclusions From OpenAI’s EU Business AI Claim, Check The Base 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 that its AI agents generated 55% of output tokens in its EU business accounts by September 30, 2026, but that figure does not mean agents handled most business AI tasks or created most of the value. The numbers, published in two reports on October 14, describe use of OpenAI products by its own ChatGPT Enterprise and Business customers, offering a detailed but limited view of workplace AI adoption in Europe.

OpenAI’s Economic Research report, Europe’s Agentic Opportunity: Moving from Breadth to Depth with AI Agents, counts output from Codex, its coding agent, and ChatGPT Work, an agent mode launched in July. By September 30, Codex generated 34% of output tokens in the covered accounts and ChatGPT Work generated 21%. Standard ChatGPT accounted for the remaining 45%.

The report also says weekly active users of Codex or ChatGPT Work among businesses with 200 or fewer users rose 8.1 times between the first week of May and the week ending October 5. OpenAI reports at least fivefold growth in every EU member state, with Belgium at 12.1 times, Italy at 10.8 times and Luxembourg at 10 times. It does not give the absolute number of users behind that growth figure.

Use varied widely across member states in OpenAI’s customer data. The agentic share of output tokens ranged 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. Those figures describe the accounts OpenAI observed; the report does not provide country-level account or user counts.

What the Token Share Shows

The 55% figure signals that agents account for a large share of generated text in OpenAI’s EU business accounts. It may help businesses and policymakers track where activity on the platform is moving as companies try tools that can work across files, browsers and other tools. It is not, on its own, evidence that most European companies have adopted agents or that agents perform most workplace AI tasks.

That distinction matters because output volume and business impact are different measures. OpenAI’s own Enterprise Signals page calls tokens an imperfect measure of business value and says agentic tasks can require substantially more computation. A workflow in which an agent reads files, plans, writes and revises can generate more text than a brief chat response, even if both count as one interaction. The report therefore offers a measure of platform activity, not a direct measure of productivity, savings or quality.

The growth statistic also needs its time window and starting point. The 8.1-times rise compares weekly active users with the first week of May, before ChatGPT Work launched. OpenAI says a major part of the increase followed that launch. The data support a claim of rapid growth in use of these products from that baseline, but without absolute user counts they do not establish how widespread use is among EU small businesses overall.

How OpenAI Defines Agent Use

The October 14 publication brought together OpenAI’s EU-focused report and its Global Affairs team’s post, The EU’s Next Chapter in AI Is Agentic. The analysis covers ChatGPT Enterprise and ChatGPT Business accounts in the 27 EU member states. Its definition of agentic use is specific to two OpenAI products: Codex and ChatGPT Work.

Product timing shapes the figures. Codex already contributed roughly a fifth of EU output tokens in May, according to the chart described in the report. ChatGPT Work was introduced in July and reached a 21% share by September 30. That means a substantial part of the reported increase in agentic output coincided with a new product becoming available, rather than measuring a longer period of change in business practice.

OpenAI also compares its highest-use customers with other firms. It defines “frontier firms” as the top 10% by output tokens per active user. In September, EU frontier firms produced about four times as many tokens per user as typical EU firms; frontier firms outside the EU produced 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 platform-based comparisons, not a census of businesses in either region.

What the Customer Data Cannot Show

Several important comparisons remain unavailable. OpenAI does not publish absolute user counts behind the 8.1-times small-business growth figure, or sample sizes for the 27 country-level results. Without those figures, readers cannot tell how much a small initial user base shaped the growth multiples or whether a small number of large accounts affected national token shares.

The report also does not establish how many tasks were completed by agents, whether they improved results, or what financial returns customers achieved. Its token measure can reflect task complexity and repeated agent work as well as adoption. The source material further notes that the national rankings for agentic share and small-business growth do not closely track one another; neither chart alone establishes a country’s overall AI readiness.

OpenAI’s findings are based on its own business customers, not a representative survey of every EU company. The figures therefore cannot answer how widely companies using other AI services have adopted agents, or whether OpenAI’s customer mix matches the wider market.

Evidence Needed Beyond OpenAI

The report’s policy conclusion is that Europe should move from broad AI adoption to deeper use of agents, through skills, worker confidence and more accessible advanced tools. The published figures are a starting point for that discussion, but testing the conclusion will require evidence beyond output tokens: absolute adoption counts, account samples, task-level measures and outcomes such as time saved or work quality.

OpenAI has not specified in the supplied material when it will release further breakdowns or comparable results. Until more detail is available, readers should treat the 55% share and the 8.1-times growth as measurements of activity among OpenAI’s customers, with their stated time windows and limits—not as estimates of agent use across Europe’s entire business economy.

Key Questions

What does OpenAI’s 55% figure measure?

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

How did OpenAI calculate the 8.1-times growth figure?

OpenAI compared weekly active users of Codex or ChatGPT Work at businesses with 200 or fewer users in the first week of May with the week ending October 5. The report does not give the absolute user totals behind that multiple.

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

No. It covers OpenAI’s business accounts in the EU, not all European businesses, and its headline measure is generated tokens rather than the number of companies or tasks using agents.

Which EU countries had the highest agentic token shares?

In OpenAI’s data, Lithuania had the highest share at 85%, followed by Cyprus at 81%. The report does not provide country-level sample sizes, so the figures should not be read as national rankings of AI readiness.

Source: ThorstenMeyerAI.com