OpenAI Says Most EU Business AI Is Now Agentic. Read The Denominator First.

🔍 Read the full analysis: OpenAI Says Most EU Business AI Is Now Agentic. Read The Denominator First. 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 the company’s ChatGPT Enterprise and ChatGPT Business accounts across the 27 EU member states by Sept. 30. The figure comes from a report published Oct. 14, but it measures text generated by OpenAI’s tools—not the share of European business tasks handled by agents or the value those tasks produced.

OpenAI Economic Research’s report, Europe’s Agentic Opportunity: Moving from Breadth to Depth with AI Agents, counts Codex, OpenAI’s coding agent, and ChatGPT Work, an agent mode launched in July. In September, Codex accounted for 34% of output tokens in the covered accounts and ChatGPT Work for 21%. The other 45% came from standard ChatGPT use.

The company also reports that weekly active users of Codex or ChatGPT Work at businesses with 200 or fewer users increased 8.1 times between the first week of May and the week ending Oct. 5. OpenAI gives no absolute user totals in the supplied report summary and says a major part of the growth followed ChatGPT Work’s introduction. Each EU member state recorded growth of at least fivefold; Belgium, Italy and Luxembourg had the largest reported multiples.

Agent use varied by country in OpenAI’s customer data, from 85% of output tokens in Lithuania and 81% in Cyprus to 22% in Slovenia and 26% in Austria. Germany’s share was 51%, placing it 17th among the 27 states. The report also says the top 10% of firms by output tokens per active user in the EU generated about four times the output of typical EU firms, while comparable frontier firms outside the EU generated more than twice the EU frontier level.

What the Token Share Shows

The figures offer a window into how customers of one major AI provider are using its products at work. They suggest that agent products account for a substantial portion of OpenAI’s generated output in its EU business accounts, and that use among its smaller business customers rose quickly over the period measured. Those signals matter to companies weighing agent tools, workforce training and access to more advanced AI capabilities.

But the headline does not establish that agents now handle most business AI work across Europe. OpenAI’s data covers its own Enterprise and Business accounts, not all firms or AI providers. Tokens are generated text, and agent workflows can use more tokens than a simple chat exchange because they may read material, plan, produce outputs and check results. A high token share can therefore reflect the computational intensity of agent runs rather than a majority of tasks or users.

The report’s policy argument is that Europe should turn broad AI adoption into deeper use by building worker skills and confidence and making advanced tools more accessible and affordable. The reported gap between EU and non-EU frontier firms gives that argument a business-productivity dimension, but token output alone does not measure productivity or prove what caused the gap.

How OpenAI Defines Agent Use

The report’s scope is limited to ChatGPT Enterprise and ChatGPT Business accounts in the 27 EU member states. It labels Codex and ChatGPT Work as agentic tools and compares their output tokens with those from plain ChatGPT. OpenAI’s global Enterprise Signals page supplies additional worldwide context, but the reported country and small-business figures describe activity within OpenAI’s customer base.

The timeline matters. Codex already accounted for roughly one-fifth of EU output tokens in May, according to the report’s chart. ChatGPT Work launched in July and contributed 21% by Sept. 30. That means the measured shift includes both increasing use of an existing agent and uptake of a newly introduced product. Small-business growth is calculated from the first week of May, before the launch.

OpenAI’s stated conclusion is that Europe has an opportunity to move from broad adoption toward more extensive use of agents. The report does not treat every business as equally advanced: its “frontier firms” are defined as the top 10% by output tokens per active user. It says EU companies make up about 10% of frontier firms worldwide, compared with about 65% in North America.

What the Data Cannot Establish

The report does not disclose country-by-country account or user counts in the supplied material, making it difficult to judge how representative each national percentage is. OpenAI itself notes that smaller countries can have high agent use within larger firms. The rankings should therefore be read as a description of OpenAI’s customers, not a definitive league table of national AI adoption.

There are also no absolute small-business user totals, task counts, or measures of business outcomes such as time saved, costs or revenue. The 8.1-times figure is a growth multiple from a May baseline; it does not show how many businesses or workers adopted agents. Nor does a 55% token share reveal what fraction of interactions were agentic. The source’s illustrative estimate that a token share might translate into a much smaller interaction share depends on an assumed difference in output per agent run and is not a report finding.

It remains unclear how much of the reported change reflects lasting changes in company workflows, the launch of ChatGPT Work, differences in customer composition, or other factors. OpenAI’s data cannot establish how business use across the EU compares with use of other AI products.

Tracking Adoption Beyond Launch

The next useful evidence would include updated figures after ChatGPT Work has been available for longer, along with absolute counts of participating businesses and active users. Further reporting on tasks, outcomes and the distribution of use across company sizes would help distinguish product uptake from changes in how work is done.

OpenAI’s report calls for more skills training and wider access to advanced tools but does not, in the supplied material, set out a specific timetable or measurable target for that policy agenda. Until comparable follow-up data is available, the 55% figure is best treated as a snapshot of output within OpenAI’s EU business accounts, not a measure of agent use across the European economy.

Key Questions

What does OpenAI’s 55% figure measure?

It is the share of output tokens generated by Codex and ChatGPT Work in OpenAI’s covered EU ChatGPT Enterprise and Business accounts by Sept. 30, 2026.

Does that mean agents handled 55% of business AI tasks?

No. The figure counts generated text, not tasks, interactions, users or business value. Agent workflows may produce more tokens than ordinary chat exchanges.

How much did small-business agent use grow?

OpenAI says weekly active users of Codex or ChatGPT Work at businesses with 200 or fewer users rose 8.1 times from the first week of May to the week ending Oct. 5. The company does not provide absolute user totals in the supplied material.

Why did agent output rise during the period?

The report’s chart shows Codex use rising, while ChatGPT Work launched in July and accounted for 21% of EU output tokens by Sept. 30. OpenAI says a major part of small-business user growth followed that launch.

Do the country figures represent all businesses in each EU state?

No. They cover OpenAI’s business accounts, and the report does not provide country-level sample sizes in the supplied material. They should not be read as comprehensive national adoption rates.

Source: ThorstenMeyerAI.com