Workers don’t know how to use AI — and companies are to blame, research finds

Summary

New research from Forrester finds an “alarming” lack of employee proficiency with AI tools. Forrester measured an artificial intelligence quotient (AIQ) across workers and says learning has largely stagnated despite wide deployment of AI pilots and licences. The firm pins responsibility on employers for not creating the right learning and engagement environment, which is now a bottleneck to productivity and return on investment.

Key Points

Forrester’s AIQ shows many workers lack practical understanding of AI tools — a trend the firm calls “alarming.”
Knowledge of prompt engineering rose only from 22% in 2024 to 26% in 2025 — a tiny improvement for a core skill for tools like Microsoft 365 Copilot.
Forrester says the problem is organisational: employers are rolling out AI without sufficient training, learning programmes or governance.
Low AIQ can mean employees misuse tools, fail to spot incorrect outputs or abandon AI altogether, undermining productivity gains.
Surveys show a gap between leadership expectations and reality: while most C-suite execs expect AI to boost output, many employees report increased workload and reduced trust when given inaccurate AI-generated work.

Content summary

Forrester measured workers’ ability to use and reason about AI using an AIQ metric and found progress is minimal despite heavy investment in AI licences and pilots. The report highlights prompt engineering as an example — only a four percentage point uptick year-on-year — and argues that this slow learning curve is preventing organisations from realising productivity benefits.

The researchers say employers are at fault: many have deployed AI tools assuming they are intuitive, then failed to provide structured training, ethical guidance or support. The result is a mixed experience where skilled users boost productivity while others either misuse tools or stop using them. Complementary studies cited in the piece show executives are broadly optimistic about AI’s potential, but employees often see increased workloads and lose trust when leaders circulate polished-but-flawed AI outputs (so-called “workslop”).

Context and relevance

This matters to HR leaders, L&D teams and executives planning AI roll-outs. Organisations keen to extract value from AI must treat adoption as a people and learning challenge, not just an IT or procurement exercise. The story ties into broader trends: growing demand for AI skills in job specs, concerns about the ethics and governance of AI at work, and the need for upskilling to protect productivity and trust.

Why should I read this?

Quick version: if your organisation has bought AI tools but hasn’t trained people properly, you’re probably wasting money and frustrating staff. This piece points that out bluntly — and gives you the ammunition to argue for real training, governance and time to build AI competence across the workforce.