OpenAI’s Claim About Agentic AI In EU Business Needs Context

🔍 Read the full analysis: OpenAI’s Claim About Agentic AI In EU Business Needs Context on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get hardware and tech essentials 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 AI agents generated 55% of output tokens in ChatGPT Enterprise and Business accounts across the 27 EU member states by September 30. The figure documents growing use of OpenAI’s agent products among its own business customers, but it does not mean agents handle most workplace AI tasks or deliver most of the business value.

OpenAI’s Economic Research report, Europe’s Agentic Opportunity: Moving from Breadth to Depth with AI Agents, groups Codex, the company’s coding agent, and ChatGPT Work, an agent mode launched in July, under the term “agentic.” The report says Codex generated 34% of EU business-account output tokens and ChatGPT Work generated 21%. The remaining 45% came from standard ChatGPT use.

For businesses with 200 or fewer users, OpenAI says weekly active users of Codex or ChatGPT Work rose 8.1 times from the first week of May to the week ending October 5. The company reports at least fivefold growth in every member state, with Belgium at 12.1 times, Italy at 10.8 times and Luxembourg at 10.0 times. It does not provide absolute user counts alongside those growth figures.

The report also finds wide variation in agentic output share by country: Lithuania was at 85% and Cyprus at 81%, while Slovenia was at 22% and Austria at 26%. Germany’s share was 51%, placing it 17th among the 27 countries. OpenAI says these are measures of activity in its accounts, not a comprehensive survey of businesses across Europe.

What the Token Share Shows

The findings offer a view of how companies using OpenAI’s paid business products are adopting its agents. They may help businesses and policymakers track whether staff are moving from asking AI questions to assigning it work involving files, browsers or company tools. But the 55% figure is a measure of generated tokens—the text models produce—not a count of interactions, users, completed tasks or financial returns.

That distinction matters because agent workflows can generate more text than a simple chat response. An agent may inspect files, plan, produce a result and revise it, using tokens at each stage. OpenAI’s Enterprise Signals page itself describes tokens as an imperfect measure of business value and says agentic tasks can require substantially more computation. A high token share can therefore indicate intensive use without showing how many employees used agents or whether the work improved productivity.

The report frames Europe’s challenge as turning broad AI adoption into deeper use, including building worker skills and making advanced tools more accessible and affordable. That is a policy argument from the company, not an outcome established by the usage statistics. For firms considering adoption, the report supplies a signal about activity within OpenAI’s customer base, but not independent evidence of returns or economy-wide uptake.

How OpenAI Defines Agent Use

The report covers ChatGPT Enterprise and ChatGPT Business accounts in EU member states, rather than all European workplaces or all AI providers. Its “agentic” category includes Codex and ChatGPT Work. Those product choices shape what the report counts and make the figures useful for understanding OpenAI’s own offerings, not for estimating the full market.

Product timing also affects the reported growth. OpenAI’s chart shows Codex at roughly one-fifth of EU output tokens in May and 34% by September 30. ChatGPT Work was introduced in July and accounted for 21% by the end of September. The report says a major part of the small-business user growth followed that launch. The increase thus combines adoption of an existing coding agent with uptake of a newly introduced product.

OpenAI also compares high-use companies with typical firms. It defines “frontier firms” as the top 10% by output tokens per active user. In September, EU frontier firms generated about four times as many tokens per user as typical EU firms, while frontier firms outside the EU generated more than twice as many as the EU frontier. The report says EU companies made up about 10% of frontier firms worldwide, compared with about 65% in North America. These comparisons use token output per user as their measure, not independent measures of AI capability or business performance.

What the Data Cannot Establish

OpenAI does not publish the account or user sample sizes for each country in the supplied report material. Without those counts, the country rankings cannot show whether a result reflects a broad pattern or a smaller group of large, heavily agent-focused customers. OpenAI notes that smaller countries including Cyprus, Estonia and Lithuania have relatively high agentic use within their larger firms.

The report also does not provide absolute small-business active-user numbers, the number of tasks completed by agents, or evidence that the tools improved output, costs or revenue. The 8.1-times growth figure uses the first week of May as its baseline, before ChatGPT Work launched; the company does not provide enough information here to establish how large the starting user base was. The source material also does not include a completed explanation of every limitation raised in its discussion, so no additional conclusions about the country comparisons should be inferred.

It remains unclear how representative OpenAI’s business customers are of EU companies overall, how the results compare with other AI providers, and whether agent usage will persist after the initial uptake of ChatGPT Work. The figures are vendor-reported observations about OpenAI products, not an independent measure of European business adoption.

What to Watch in Future Data

The next useful evidence would include updated usage figures with absolute user and account counts, clearer country-level sample sizes and consistent time periods. Independent research measuring completed tasks, productivity, costs and user outcomes would help determine whether increased agent activity translates into business gains.

OpenAI says Europe should expand worker skills and access to advanced AI, but the report does not set out a measured timetable for those changes or a specific next data release. Until further evidence is available, the October figures are best read as a snapshot of agent use among OpenAI’s EU business customers, with the effects on the wider economy still uncertain.

Key Questions

What does OpenAI mean by “55% of output tokens”?

It means Codex and ChatGPT Work generated 55% of the text tokens produced in the covered EU ChatGPT Enterprise and Business accounts by September 30. It is not a claim that agents handled 55% of users’ tasks or generated 55% of business value.

Which products count as agents in the report?

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

Did small-business agent use rise 8.1 times across Europe?

OpenAI reports an 8.1-times increase in weekly active users of Codex or ChatGPT Work among businesses with 200 or fewer users, comparing the first week of May with the week ending October 5. The report does not give the absolute number of users, and says a major part of the growth followed ChatGPT Work’s launch.

Do the country rankings measure national AI readiness?

No. They describe activity in OpenAI’s accounts, and the report material does not give country-level account or user sample sizes. The rankings should not be treated as a full comparison of AI use or readiness across EU economies.

Does the report show that agents increase productivity?

No. The report measures product use and token output; it does not establish productivity gains, financial returns or improved business outcomes. Those questions require evidence beyond usage totals.

Source: ThorstenMeyerAI.com