The Human Problem Behind Every Failed AI Investment

A recent study from Pearson and AWS offers an uncomfortable answer: The workforce isn’t ready.

The study surveyed 2,711 employers, higher education leaders and students across six countries.

Only 14% of graduates say they can apply AI effectively in real workflows. Among employers, 53% rank AI skills as their single biggest hiring challenge. The weakest competency recruiters find in candidates isn’t technical. It’s critical evaluation.

The confidence gap is striking. More than three in four higher education leaders believe their graduates already meet employer expectations. The data disagrees.

BCG found that while 75% of C-suite leaders rank AI among their top three priorities, only 25% say their organizations are realizing significant value. The explanation may be structural: BCG attributes just 10% of AI value creation to algorithms and 20% to infrastructure. The remaining 70% hinges on people, processes and change management: the hardest parts to scale and the ones most organizations are underinvesting in.

The Education-to-Workforce Disconnect

The Pearson and AWS study found that employers rank communication and collaboration at 50% and adaptability at 45% as their highest priorities for graduates entering AI-enabled roles. Higher education leaders undervalue communication and collaboration by 15 percentage points relative to employer demand, and adaptability by nine points. In the U.K., a 27-point gap separates what universities prioritize from what employers ask for.

Across degree programs and markets, the study found that students are less convinced that a hybrid skillset of technical and human skills is expected. Employers are the least likely to expect that human skills alone are sufficient for an AI-driven economy.

The IMF finds that about 1 in 10 job vacancies in advanced economies now requires at least one new AI-related skill, and those positions command higher wages. But the diffusion of AI cuts both ways: in occupations with high AI exposure and little room for distinctly human contribution, employment is falling. The supply of workers who can fill the roles where it pays to know AI is not keeping up with demand.

Where ROI Stalls Inside the Enterprise

PYMNTS Intelligence found that more than seven in 10 senior technology executives at large U.S. firms say internal constraints are holding back AI performance within their organizations. Only 11% blame the technology itself.

Data quality, budget constraints and governance processes were the three most common bottlenecks, cited by 63%, 49% and 48% of respondents respectively, according to PYMNTS Intelligence.

“Companies launch lots of AI pilots but can’t turn them into repeatable, scalable value,” said Allison Bailey, BCG senior partner and global co-lead of BCG U. “There’s too much emphasis on the tech and not enough on skills development for employees”.

Another PYMNTS Intelligence survey of 60 CFOs at U.S. firms generating at least $1 billion in annual revenue found that 71% of services firms cited AI skill gaps as a key barrier. Only 12% of CFOs overall considered their companies very prepared for the workforce impact of AI.

How Enterprises Are Responding

Deloitte reported that 53% of organizations are educating the broader workforce to raise AI fluency, 48% are designing upskilling and reskilling strategies and 36% are hiring specialized talent to drive AI initiatives. According to Deloitte, most successful organizations reimagine jobs to combine human strengths and AI capabilities, with new roles including AI operations managers, human-AI interaction specialists and quality stewards.

BCG found that only 36% of employees are satisfied with their AI training, even though nearly three in four are already using AI regularly. Additionally, BCG reported that creating persona-based learning journeys drives employee AI adoption 20 times higher than a broad-based approach.

For all PYMNTS AI and digital transformation coverage, subscribe to the daily AI and Digital Transformation Newsletters.