How EU Businesses Are Adopting Agentic AI

Peter Bubenik ยท OpenAI ยท ยท Source

Step-by-Step Study Material

Step 1: What Is Agentic AI?

Basic AI Use (What came before)

Users ask questions โ†’ AI answers

Agentic AI Use (The new era)

Users assign tasks โ†’ AI plans AND completes them autonomously

Key Examples Mentioned

ToolPurpose
CodexLikely coding and technical task automation
ChatGPT WorkWorkplace task execution end-to-end

โœ… Check Your Understanding

Can you explain in one sentence why agentic AI represents a fundamentally different relationship with AI tools?


Step 2: Understanding the Growth Data

EU Small & Medium Businesses (SMBs)

  • Defined as businesses with 200 or fewer users
  • Agentic weekly active users grew 8.1 times since May 2026
  • Growth occurred across every EU member state

EU Enterprise Customers

  • Codex and ChatGPT Work now account for more than half of all output tokens
  • This means agentic use is already the dominant mode of AI interaction

๐Ÿ’ก Key Concept: Output Tokens

Output tokens = the volume of content AI generates More tokens = more intensive, complex AI use

โœ… Check Your Understanding

Why is measuring output tokens a useful way to compare AI usage intensity between companies?


Step 3: Recognizing Uneven Adoption

The adoption is NOT uniform across the EU

Slovenia โ†’ 22% of tokens from agentic tools  (LOW)
Lithuania โ†’ 85% of tokens from agentic tools (HIGH)

What This Tells Us

  • Different countries are at very different stages of agentic adoption
  • The type of work brought to AI also varies by country
  • This creates competitive inequality within the EU itself

โœ… Check Your Understanding

What factors might explain why Lithuania uses agentic AI far more intensively than Slovenia?


Step 4: The Frontier Firm Gap

Defining Frontier Firms

Frontier firms = Top 10% of companies by output tokens per active user

Three-Level Comparison

GLOBAL FRONTIER FIRMS
        โ†‘
    (2x more than EU frontier)

EU FRONTIER FIRMS
        โ†‘
    (4x more than typical EU firms)

TYPICAL EU FIRMS

The Structural Problem

  • EU frontier firms do lead typical EU firms significantly
  • BUT global frontier firms still outpace EU frontier firms by 2x
  • EU firms represent only 10% of global frontier firms
  • This is disproportionately low given the EU's share of registered companies worldwide

โœ… Check Your Understanding

Why is it significant that EU firms make up only 10% of global frontier firms despite the EU's large share of registered companies?


Step 5: Implications and the Path Forward

The Core Opportunity

The article argues the EU must convert:

BROAD ADOPTION โ†’ DEEPER USE
                      โ†“
              Stronger Businesses
                      โ†“
              Higher Productivity
                      โ†“
              Benefits for Workers

Three Recommended Actions

  1. Help businesses move beyond experimentation into real operational use
  2. Give workers skills and confidence to use AI effectively
  3. Make advanced capabilities more accessible and affordable

โœ… Check Your Understanding

Why is "broad adoption" alone insufficient? What is the difference between experimenting with AI and deeply integrating it?


Summary Table

ConceptKey Fact
Agentic AI definitionAI that executes tasks, not just answers questions
SMB growth8.1x increase in weekly active users since May 2026
Enterprise share50%+ of output tokens now from agentic tools
Country range22% (Slovenia) to 85% (Lithuania)
EU vs global frontierGlobal frontier firms generate 2x more tokens than EU frontier firms
EU share of global frontierOnly 10%

Final Reflection Question

"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.