OpenAI report shows agentic AI output surges in European SMBs
New data reveals AI agents now drive over half of enterprise token output in the EU, though growth lags behind North America.
Key points
- AI agents account for 55% of output tokens for OpenAI enterprise customers in the EU.
- European small and medium businesses increased agent usage more than eightfold in under six months.
- The EU represents only 10% of global "frontier firms," trailing significantly behind North America.
OpenAI released a report detailing a significant shift in artificial intelligence usage across European enterprises. According to the data, agentic AI now accounts for more than half of all output tokens generated by the company’s business customers in the European Union. This marks a transition from simple question-answering tools to systems capable of completing complex tasks autonomously.
The metrics rely on output tokens rather than the percentage of companies adopting the technology. This distinction means the report highlights increased intensity of use among adopters rather than a wider spread of users. Basic generative AI handles the remaining 45% of output. The surge in agentic activity coincides with the launch of ChatGPT Work, a tool integrated into OpenAI’s broader platform.
Disparities in Regional Adoption
Adoption rates vary considerably across individual European nations. Lithuania leads with 85% of output attributed to agents, followed by Cyprus at 81% and Hungary at 79%. In contrast, Slovenia records only 22%, while Austria sits at 26% and Portugal at 29%. These figures illustrate that the transition to agentic workflows is not uniform across the continent.
Small and medium-sized enterprises are driving this change. OpenAI observed an eightfold increase in agent usage among European SMEs in less than six months. The most significant spike began when ChatGPT Work became publicly available. This suggests that smaller businesses are leveraging accessible tools to implement capabilities previously reserved for larger organizations with deeper technical resources.
Global Context and Growth Rates
Despite the rapid local growth, Europe trails other global regions in high-level AI adoption. The European Union accounts for approximately 10% of the world’s "frontier firms." These are companies ranking in the top tenth of businesses by AI output tokens per active user. North America dominates this category with 65% of global frontier firms.
The Asia-Pacific region holds 13% and the EMEA region 11%. Growth momentum also differs significantly. Non-EU frontier firms experienced a 12.9-fold growth in output between May and September. EU firms saw a more modest 6.1-fold increase during the same period. OpenAI CFO Sarah Friar stated that Europe is entering a new era of AI with task-carrying tools.
What should you do?
Security and technology professionals should monitor how agentic AI tools interact with internal systems. As these agents perform tasks autonomously, their access permissions and operational boundaries require careful definition. Organizations should assess whether their current governance models account for AI-driven actions rather than just human-initiated queries.
Background: Generative AI
Generative AI examples include text completion, image synthesis and code generation. These systems predict the next likely token based on vast training datasets. Understanding the underlying mechanics helps you identify risks like data leakage and hallucination in your infrastructure.
Read the full guide: Generative AI Examples: Mechanisms, Risks and Control
What to do and how to stay safe: OpenAI
- Review access permissions granted to AI tools to ensure they align with the principle of least privilege.
- Monitor logs for unusual volumes of automated output or unexpected task completions from AI agents.
- Evaluate vendor reports on AI usage to understand industry benchmarks and potential security implications.
- Establish clear policies distinguishing between human-operated AI queries and autonomous agentic tasks.
Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation
General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
What percentage of EU enterprise output tokens come from agentic AI?
According to the OpenAI report, agentic AI accounts for 55% of output tokens among enterprise customers in the EU.
How does European growth compare to other regions?
Non-EU frontier firms saw 12.9x growth between May and September, while EU firms grew by 6.1x in the same period.
Which European countries have the highest agentic AI usage?
Lithuania leads with 85%, followed by Cyprus at 81% and Hungary at 79%, according to the report's data.
