Users ask questions โ AI answers
Users assign tasks โ AI plans AND completes them autonomously
| Tool | Purpose |
|---|---|
| Codex | Likely coding and technical task automation |
| ChatGPT Work | Workplace task execution end-to-end |
Can you explain in one sentence why agentic AI represents a fundamentally different relationship with AI tools?
Output tokens = the volume of content AI generates More tokens = more intensive, complex AI use
Why is measuring output tokens a useful way to compare AI usage intensity between companies?
Slovenia โ 22% of tokens from agentic tools (LOW)
Lithuania โ 85% of tokens from agentic tools (HIGH)
What factors might explain why Lithuania uses agentic AI far more intensively than Slovenia?
Frontier firms = Top 10% of companies by output tokens per active user
GLOBAL FRONTIER FIRMS
โ
(2x more than EU frontier)
EU FRONTIER FIRMS
โ
(4x more than typical EU firms)
TYPICAL EU FIRMS
Why is it significant that EU firms make up only 10% of global frontier firms despite the EU's large share of registered companies?
The article argues the EU must convert:
BROAD ADOPTION โ DEEPER USE
โ
Stronger Businesses
โ
Higher Productivity
โ
Benefits for Workers
Why is "broad adoption" alone insufficient? What is the difference between experimenting with AI and deeply integrating it?
| Concept | Key Fact |
|---|---|
| Agentic AI definition | AI that executes tasks, not just answers questions |
| SMB growth | 8.1x increase in weekly active users since May 2026 |
| Enterprise share | 50%+ of output tokens now from agentic tools |
| Country range | 22% (Slovenia) to 85% (Lithuania) |
| EU vs global frontier | Global frontier firms generate 2x more tokens than EU frontier firms |
| EU share of global frontier | Only 10% |
"The EU's opportunity is to turn broader adoption into deeper use."
Based on everything you have learned, write a short paragraph explaining what this statement means and what obstacles might prevent the EU from achieving it.