OpenAI Economic Research's latest report shows that across the European Union businesses are shifting from question-based AI use to agentic tools that plan and execute tasks. The analysis draws on data from ChatGPT Enterprise and Business accounts and examines how agentic adoption is evolving, with emphasis on usage intensity, the types of work handled by AI, and differences between countries and firm types.
Key findings and figures
- In the EU sample, Codex and ChatGPT Work together now account for more than half of output tokens.
- Small and medium businesses (SMBs) are defined in the report as businesses with 200 or fewer associated users.
- The agentic shift includes SMBs: since May 2026 weekly active users of Codex and ChatGPT Work among EU SMBs have grown more than eightfold, with increases observed in every member state.
Large variation across countries and use cases
- Adoption of agentic tools is uneven across the EU. Among ChatGPT Enterprise accounts in the sample, the share of output tokens generated through Codex and ChatGPT Work ranged from 22% in Slovenia to 85% in Lithuania.
- Differences go beyond intensity: the kinds of work organizations bring to ChatGPT also vary by country.
Frontier firms and international comparisons
- The report defines frontier firms as the top 10% of firms by output tokens per active user. The sample includes only ChatGPT Enterprise and Business accounts.
- EU frontier firms use AI far more intensively than typical EU firms: in September, frontier firms in the EU generated more than four times as many output tokens per active user as typical EU firms.
- Despite that, frontier firms outside the EU generated more than twice as many output tokens per active user as frontier EU firms.
- EU-based firms made up only 10% of frontline (frontier) firms globally, a share that is low relative to the EU's larger share of registered companies worldwide.
Implications and next steps
These figures point to a significant gap in how intensively businesses deploy AI. As agentic capabilities improve, the EU’s opportunity is to convert broader adoption into deeper usage that strengthens firms, boosts productivity, and yields tangible benefits for workers. Closing that gap will require moving organizations beyond experimentation, equipping workers with the skills and confidence to use AI effectively, and making advanced capabilities more accessible and affordable.
Conclusion
The OpenAI Economic Research analysis documents a clear move toward agentic AI across both large enterprises and SMBs in the EU, but also highlights wide disparities in intensity and application. To realize the potential productivity gains, policymakers and businesses will need to focus on deeper integration, workforce upskilling, and improving access to advanced AI capabilities.
