OpenAI Says Most EU Business AI Is Agentic. Here’s What To Ask Next

🔍 Read the full analysis: OpenAI Says Most EU Business AI Is Agentic. Here’s What To Ask Next 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 Codex and ChatGPT Work generated 55% of output tokens in its ChatGPT Enterprise and Business accounts across the 27 European Union member states by September 30, 2026. The figure, in a report published October 14, describes the company’s own customers and measures text generated by its models—not the share of 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 at 34% and ChatGPT Work at 21% of output tokens in the covered EU accounts. The remaining 45% came from what the report classifies as plain ChatGPT. OpenAI defines “agentic” use here as activity with Codex, its coding agent, or ChatGPT Work, an agent mode that can work with files, browsers and company tools.

The report also says that among businesses with 200 or fewer users, weekly active users of Codex or ChatGPT Work rose 8.1 times between the first week of May and the week ending October 5. OpenAI says every EU member state recorded growth of at least five times over that period; Belgium, Italy and Luxembourg had the largest reported increases. The report does not give absolute user counts for this measure.

Reported agentic output shares varied substantially by country, from 85% in Lithuania and 81% in Cyprus to 22% in Slovenia and 26% in Austria. Germany’s share was 51%, which OpenAI’s country ranking placed 17th among the 27 member states. These figures describe activity in OpenAI accounts, not national business use as a whole.

What the Token Share Can Show

The report offers a timely view of how quickly businesses using OpenAI’s products are trying agent features. If the trend continues, it could affect how companies plan software access, staff training and spending on AI tools. But the headline percentage is narrower than “most business AI is agentic”: it refers to tokens generated in OpenAI’s EU business accounts, not all European firms, users or work tasks.

Tokens are units of model input and output, and agent workflows can generate more output than a short chat response because they may read material, plan, produce work and revise it. OpenAI’s Enterprise Signals materials describe tokens as an imperfect measure of business value and note that agent tasks can require substantially more computation. A larger token share could reflect longer, more compute-intensive sessions, rather than agents being used for most interactions or delivering more economic benefit.

The 8.1-times growth figure also needs its starting point. It compares weekly active users with the first week of May, before ChatGPT Work launched in July. OpenAI says a major part of the increase followed that launch. The result signals rising use of these products from that baseline; without absolute counts, it cannot establish how widely agents have spread among small EU companies overall.

A New Product Shapes the Comparison

The report combines data from ChatGPT Enterprise and ChatGPT Business accounts in the EU’s 27 member states. Its September snapshot reflects both Codex and ChatGPT Work. The latter launched in July 2026, so it was not present at the beginning of the May-to-October period used for the small-business growth comparison.

OpenAI’s charts show Codex at roughly one-fifth of EU output tokens in May and 34% by September 30, while ChatGPT Work accounted for 21% after its launch. That timing matters when interpreting the increase: part of the shift reflects the arrival of a product included in the report’s definition of agentic use. OpenAI’s stated policy conclusion is that Europe should turn broad AI adoption into deeper use by developing workers’ skills and confidence and making advanced capabilities more accessible and affordable.

The report also compares high-use companies with typical firms. It defines “frontier firms” as the top 10% by output tokens per active user. OpenAI says EU frontier firms generated about four times the output tokens per user of typical EU firms in September, while frontier firms outside the EU produced more than twice as many as the EU frontier. Those are measures of output intensity, not a direct comparison of productivity or business returns.

What OpenAI’s Data Leaves Out

The report does not provide country-level account or user sample sizes, absolute weekly active-user numbers for the small-business comparison, or a breakdown of how many tasks were completed with agents. It also does not measure time saved, quality, revenue or other business outcomes. Those omissions make it difficult to judge how representative the country rankings are or whether higher token use corresponds to higher value.

Because the figures come from OpenAI accounts, they cannot establish the rate of agent adoption across all EU businesses or across other AI providers. The report’s country comparisons may also reflect differences in which customers use OpenAI products, including the presence of a few large or highly agent-intensive accounts. The data do not settle whether the reported growth is a lasting shift in business practice or a response to a new product launch.

Evidence Needed Beyond Usage

The next useful evidence would include absolute user and account counts, consistent measures before and after ChatGPT Work’s July launch, and country-level sample sizes. Independent or company-reported measures of task completion, quality, time saved and cost would help distinguish greater model activity from greater business value.

OpenAI has framed the report as a case for moving from broad adoption to deeper use. Whether that happens will depend on how companies deploy agents and train workers, as well as whether the tools produce reliable results at a justifiable cost. The October report provides a snapshot of activity in OpenAI’s EU business accounts; it does not yet answer those questions.

Key Questions

What does OpenAI mean by “agentic” use?

In this report, agentic use means activity with Codex or ChatGPT Work. OpenAI describes ChatGPT Work as an agent mode that can use files, browsers and company tools.

Does the 55% figure mean most EU businesses use AI agents?

No. It is the share of output tokens in OpenAI’s covered EU Enterprise and Business accounts. It does not measure the share of businesses, users or tasks using agents, and it does not cover all EU AI use.

How did small-business agent use change?

OpenAI reports that 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 October 5, 2026. The report gives no absolute user totals, and part of the growth followed ChatGPT Work’s July launch.

Does the report show that agents improve business performance?

No. It reports product usage and output-token shares, not measured effects on productivity, costs, revenue or work quality.

Source: ThorstenMeyerAI.com