AI investment is booming, with businesses around the world pouring unprecedented sums into the technology in the hopes of squeezing out a competitive edge, but new figures have revealed that very few have managed to translate their AI ambitions into repeatable, scalable value.

Surveying more than 630 senior executives at enterprise firms across London, New York, Singapore, Sydney and Tokyo, the Economist Impact’s latest report, The Leaders’ Guide to Building AI-Powered Workplaces, found that while 88% of firms see AI as offering a strategic advantage, just 4% have turned early AI pilots into engines of growth.

The think tank found that most execs (73%) are focused on quick productivity wins, with nearly eight in ten measuring success through employee productivity metrics, while far fewer track long-term outcomes such as engagement, skills development, or retention. 

The data shows productivity is still the dominant measure of AI value across global markets. In Tokyo, for example, 88% of leaders said that employee productivity is the top ROI metric for their AI investments, more than any other city surveyed, but not too far above New York (72%), Singapore (78%), or Sydney (69%).

However, despite employee performance being the top priority around the world, and 99% of executives claiming their firm has a strategy for developing AI-relevant skills, only 38% have dedicated, sufficient budgets for AI-related training.

The study found that most businesses currently rely on informal or ad hoc AI training, with mentorship (54%) and self‑directed online courses (52%) being the most popular, while external partnerships (21%) and structured internal training (16%) are by far more uncommon. 

Further, the data shows a mismatch between the skills executives believe to be most important and where AI training is actually focused.

For instance, leaders rate critical thinking and creativity just as important as technical expertise (both 95%), yet only around a third believe their employees are strong in all of these areas, a gap that limits organisations’ ability to adapt to change and push innovation

A diffusion of responsibility is further weakening AI talent strategies, with nearly half of executives saying managers hold minimal responsibility for developing employee AI skills, while 8% report no responsibility at all. 

Resistance to change among employees and middle management is cited by one in three leaders as a key barrier to aligning talent strategy with AI goals, preventing initiatives from reaching the frontline and delivering value.

Recommended reading

• 79% of Firms Betting On AI To Raise Revenue, Only 1 in 4 See How
• AI is Rendering The Way We Measure Productivity Obsolete
• AI Hasn’t Killed Overwork – It’s Making it Worse

Where training does exist, almost half of executives say it reaches less than 10% of the workforce, restricting AI capability to small specialist teams, which the Economist Impact argues is evidence that firms are “ticking the box”, rather than building AI capability at organisational scale.

Better news for the UK, however, is that London emerges as an outlier in prioritising talent pipelines, with 53% of executives citing university partnerships as a top strategy, compared with a global average of 43%. The capital also shows stronger scaling progress, with 41% of firms expanding successful AI initiatives across multiple business units.

“Our research indicates that many businesses are still laying the tracks while the AI train is in full motion,” said Charles Ross, Asia head of policy and insights at Economist Impact. 

“Driven by the desire for quick, tangible wins, organisations often equate productivity gains with ROI. Without clear strategy, skills, and governance, they risk missing the sustainable competitive advantage AI can deliver.”