Ajelix Data: 50% of U.S. Workers Use Advanced AI Features vs 37% in the EU

What the Data Actually Shows

Ajelix, an agentic AI productivity platform with 340,000 users worldwide, analyzed anonymized behavioral data from 3,000 spreadsheet AI users between February 2025 and June 2026. The findings draw on over 48,000 spreadsheet generations and more than 32,000 user messages across U.S. and EU cohorts.

The headline figure: 50.1% of U.S. users engaged with advanced AI features—script generation, debugging, or optimization—compared with 37.3% of EU users.

The breakdown by feature makes the gap more concrete:

The debugging and multi-feature combination gaps are particularly sharp. U.S. users are more than twice as likely to debug scripts or chain multiple AI capabilities together in a single workflow.

The One Area Where EU Users Lead

The picture is not uniformly one-directional. EU users were more likely to use AI for formula generation: 66.9% did so, compared with 61.2% of U.S. users.

This suggests EU professionals are actively using AI—but concentrating usage on discrete, well-defined tasks rather than extending it into code generation or end-to-end workflow automation. The pattern points to a difference in depth of adoption, not simply a lag in awareness.

Prompt Complexity as a Behavioral Signal

Ajelix also analyzed prompt length across 32,331 user messages. The median U.S. prompt contained 15 words; the EU median was 13 words. More telling: 12.0% of U.S. prompts exceeded 150 words, versus 9.7% in the EU.

Prompt length is not a proxy for AI skill. But it does reflect task complexity. Longer prompts typically carry more context, constraints, and multi-step instructions—the kind of input associated with agentic workflows rather than single-question lookups.

As Agnese Jaunosane, COO of Ajelix, put it: U.S. professionals are “giving [AI] context, combining tools, generating code, and turning interactions into completed business assets.” That framing—AI as a production tool rather than a reference tool—captures the behavioral shift the data reflects.

How This Fits the Broader Adoption Picture

The Ajelix findings align with independent research. A 2026 Brookings study found that 43% of U.S. workers used AI in their jobs versus 32% of European workers, with U.S. workers also spending a larger share of working hours on AI tasks. A 2026 Gallup study found that frequent AI users—roughly 30% of the workforce—apply AI to coding and automation three times more often than occasional users.

The pattern that emerges across these data sources is consistent: higher frequency of AI use correlates with adoption of more advanced features. Workers who use AI daily are more likely to move from formula assistance toward script generation, debugging, and workflow automation.

Methodological Boundaries Worth Noting

Ajelix is transparent about the scope of its analysis. The data reflects behavior among its own user base—not a representative sample of all U.S. or EU workers. Feature availability was identical in both regions, which rules out access as a confounding variable. But self-selection effects, industry composition, and organizational context are not controlled for.

The findings are directionally useful, not definitively conclusive. They benchmark real-world usage patterns on a specific platform, which is more grounded than survey-based self-reporting—but should be read as one data point in a larger picture.

The Practical Takeaway

For teams evaluating AI tool adoption, the Ajelix data offers a concrete reference point: the gap between basic and advanced AI usage is not theoretical. It shows up in measurable feature engagement, prompt complexity, and the types of outputs users are generating.

If your team is primarily using AI for formula lookups or isolated Q&A, the data suggests there is significant headroom—script generation, debugging, and workflow automation are where productivity compounding tends to accelerate. The question is not whether those features exist, but whether your workflows are structured to use them.