EY Finds Firms Lose 40% of AI Gains Through Poor Training

Companies are failing to capture up to 40% of potential productivity gains from AI due to inadequate talent strategies and weak workforce foundations, according to research from EY.

The EY 2025 Work Reimagined Survey questioned 15,000 employees and 1,500 employers across 29 countries, revealing a gap between the state of AI adoption and the human readiness required to extract full value from the technology. The findings show that while 88% of employees now use AI at work, most applications remain limited to tasks such as search and document summarisation.

The research reveals that only 5% of employees are using AI in ways that fundamentally transform their work. This limited uptake of advanced applications occurs despite mounting pressure on workforces, with 64% of surveyed employees reporting perceived increases in their workloads over the past year.

EY research shows training gaps and skill erosion concerns

The survey identifies several factors contributing to unrealised productivity potential. Just 12% of employees receive what they consider sufficient AI training to unlock the technology's full benefits. Meanwhile, 37% of employee respondents express concern that overreliance on AI could erode their existing skills and expertise.

Shadow AI remains widespread across organisations despite employer attempts to provide internal tools. Between 23% and 58% of employees surveyed across various sectors globally are bringing their own AI solutions to work, indicating a disconnect between corporate provisions and workforce needs.

Kim Billeter, EY Global and EY Americas People Consulting Leader, says: “AI is everywhere – but companies seem to be missing out on its full potential, held back by a gap between adoption and human readiness. Most employees surveyed are still using AI for basic tasks, while concerns around job security, skill erosion and rising workloads are creating resistance. When organisations master both talent and technology, AI helps deliver outsized results, but neglecting the human side can erode those gains.”

EY identifies talent advantage as differentiator for AI success

The research introduces the concept of “Talent Advantage”, which describes organisations that effectively integrate talent and technology strategies. However, only 28% of organisations are currently on track to achieve this integration, according to the survey findings.

When AI adoption occurs on what EY terms “fragile talent foundations” – characterised by weak culture, ineffective learning programmes and misaligned reward structures – the potential benefits diminish substantially. The survey identifies five areas where tensions emerge between human factors and AI integration: AI adoption excellence, learning, talent health, organisational culture and reward structures.

The report also found leadership a critical factor in bridging the adoption gap. In organisations that have adopted AI effectively, 75% of employees report that their leaders are aligned on a clear AI vision. Leaders who demonstrate care, trust and empowerment drive 44% of talent health, according to the research.

EY survey reveals training paradox for AI-skilled workers

The research uncovers a retention challenge linked to AI training investment. Employees who receive more than 81 hours of annual AI training report productivity gains averaging 14 hours per week, well above the median of eight hours. However, these highly trained employees are 55% more likely to leave their organisation, as external opportunities in the competitive AI talent market often outpace internal promotion cycles.

EY suggests that employers can address retention challenges through comprehensive total rewards packages, including access to technology, flexibility and career opportunities that leverage AI skills.

The survey also documents a shift in broader talent health metrics. Global talent health, measured as employee “net promoters”, rose by 10 points year-on-year, moving from 55% to 65%: an 18% increase. 

Quit intent has fallen to 29%, the lowest level reported in four years and substantially below the 43% peak recorded during 2021’s ‘Great Resignation’. The cooling job market appears to have prompted employees to remain in their current roles, even as AI-related concerns persist.

Kim says: “The widespread adoption of AI is evident, yet many organisations are still seeing only modest returns. Our findings highlight the urgent need to address the human side of AI adoption. As AI reshapes the workplace, leaders must build cultures that support both talent health and effective technology use. Organisations that strengthen their talent foundations while advancing AI applications will be best positioned to achieve transformational results. It's about creating the right conditions for people and technology to thrive.”