The latest study on AI impact, involving nearly 6,000 executives from four countries, offers a reality check on AI’s influence in business. Contrary to the hype, AI has led to modest productivity shifts and little change in employment over the last three years. Rather than a tech failure, it's the early deployment phase at play.
Widespread, Yet Incremental
AI use is widespread with 69% of firms incorporating it, notably through large language models and machine learning. The UK saw adoption rise from 61% to 71% in 2025. Still, the numbers tell a different story about impact, with over 90% of firms reporting no measurable change in headcount due to AI. This pattern aligns with how transformative technologies develop, slowly and in phases.
Executives anticipate a 1.4% productivity increase and a 0.8% output rise in the next three years. The US expects a 2.25% gain, while the UK anticipates 1.86%. These numbers sound promising for economies battling sluggish productivity, but are they setting themselves up for disappointment?
The Employment Conundrum
jobs, executives project a 0.7% decrease in headcount, largely through slower hiring. History shows automation doesn’t necessarily mean job losses. Instead, it creates new roles, think data governance and AI service development. But are companies prepared to manage this transition smoothly?
The study shows a gap between executive and employee expectations. Executives foresee a 1.2% reduction in employment, while employees expect a 0.5% increase. Such difference arises from executives focusing on cost and competition, while employees see task augmentation. The reality is, AI often augments rather than replaces workers, especially in knowledge-intensive roles.
Bridging the Expectation Gap
Survey design affects results, with various reports on AI adoption showing differing figures. While a McKinsey survey suggests 88% adoption, the National Bureau of Economic Research reports 69%. Differences in sampling and question framing add to the disparity.
Executives expect significant shifts once AI integrates more thoroughly into workflows. The question isn’t if AI will influence productivity and employment, but how fast organizations can turn adoption into tangible economic gains. Are we underestimating the time required to achieve these gains?
AI’s transformative potential is undeniable, but the pace matters. Patience and strategic planning could be the real game-changers here. As we strip away the marketing, it’s clear: the architecture matters more than the parameter count.