AI skills training is spreading faster than the time to do it. Three surveys published between 29 September and 5 October 2026 by PwC, Workera and Workday agree that AI use at work is climbing while employer support for learning stalls or falls. For L&D and HR teams planning 2027 budgets, the gap now has numbers.
Key facts
- PwC’s survey of 49,364 workers in 48 countries and regions found access to learning and development resources fell to 51%, from 59% a year earlier.
- In the same survey, daily use of generative AI rose from 14% to 22%.
- Workera’s poll of 1,000 US employees at large firms found AI-specific training on offer at 58% of companies, up from 25%, yet 56% said they get no work hours for skill-building.
- Workday found 79% of 6,000 employees know which skills they need, but only 66% say their employer helps them build them.
AI skills training: three surveys, one pattern
| Survey | Sample | Learning finding | AI finding |
|---|---|---|---|
| PwC Global Workforce Hopes and Fears 2026 (released 29 September) | 49,364 workers, 48 countries and regions, fieldwork May–June 2026 | Access to L&D resources down from 59% to 51% | 64% used AI at work in the past year, up 10 points |
| Workera (published 29 September) | 1,000 US salaried staff at firms with 5,000+ employees, July 2026 | 56% have no work-hours time to learn; 84% spend five hours a week or less on skills | Use of AI several times a week up from 40% to 68% |
| Workday Global Workforce Report (5 October) | 6,000 full-time employees, September 2026, plus hiring data | 13-point gap between knowing needed skills (79%) and getting employer help (66%) | Demand for hands-on AI skills up 51% in job postings |
The methods differ, so the percentages cannot be added together. Workera’s is a US sample of big companies. PwC’s is global. Workday mixes surveys with data from about 550 employers using its recruiting software. Yet all three point the same way: workers are using AI more, and the share who feel supported in learning has not kept pace.
One discrepancy deserves a flag. Workera’s announcement headline says 65% of employees have no time at work to build skills, while the detailed findings published by Human Capital Leadership Review give 56%. We use the lower figure.
Is prompting still the skill to teach?
Workday’s hiring data suggests the target has moved. Demand for basic AI skills such as simple prompting fell 25% after peaking in January 2026, according to the Workday report. Demand for hands-on skills, building AI tools and automating workflows, rose 51% between September 2025 and July 2026. Mentions of training skills in job requisitions dropped 13% over the same period.
Many corporate AI courses launched in 2025 were introductions. A course that teaches staff to write a good prompt now covers what employers increasingly take for granted. The next layer is applied: redesigning a team’s workflow, checking output, deciding what not to automate. That kind of learning takes hours of practice on real tasks, which is the resource workers say they lack.
PwC’s global workforce leader Pete Brown put the risk in plain terms. Firms should keep developing their most AI-savvy staff, he said, “but not at the expense of the rest of the workforce.” The PwC survey places 56% of workers in what it calls the “engine room”, and Brown said more than half of workers are not yet getting the same benefit from AI and skills as the front-runners.
Europe already measured this problem
EU statistics reached a similar conclusion before generative AI took off. In the 2022 Adult Education Survey, 39.5% of adults aged 25 to 64 had taken part in learning in the previous 12 months, against an EU target of 60% by 2030, according to the Commission’s ESDE 2024 review. Among people who wanted to learn more, the most cited obstacle was the training schedule, at 25.3%. Cost came third, at 15.2%.
So time, more than money, was already the main brake. Employed people do better, with 47.2% in job-related non-formal learning, but adding AI to existing job demands without freeing hours repeats an old mistake. 3LNews has written about training built into the moment of work and why digital skills development needs a place on the calendar.
What this means for L&D and HR teams
- Budget hours as well as licences. Put protected learning time into 2027 plans as a line item, for example a fixed weekly block per team, and track whether it is used.
- Retire the basics course. Replace general prompting modules with short projects in which staff automate or redesign one real task and present the result.
- Train managers to coach. Workera found half of employees say they get enough coaching to advance; managers who review AI-assisted work can close part of that gap.
- Record what people learn. With internal moves down at 57% of employers in Workday’s data, documented and verifiable skills give staff a way to show readiness for new roles.
FAQ
Is AI skills training increasing?
Yes. Workera found the share of large US companies offering AI-specific training rose from 25% to 58% in about a year.
Why do employees say they cannot learn AI at work?
Mainly time. Over half of Workera’s respondents said they get no work hours for skill-building, and EU data shows schedules were already the top barrier in 2022.
Which AI skills are employers now hiring for?
Workday’s job data shows demand shifting from basic prompting toward building AI tools and automating workflows.
Sources: PwC; Benefits and Pensions Monitor; Workday; Vision Monday; Human Capital Leadership Review (Workera); European Commission, ESDE 2024.
