Prompting is becoming a baseline; hiring needs evidence of applied AI work
Workday reports that demand for basic AI skills fell after a January peak while demand for building and automation skills rose. The useful response is a work-sample ladder, not a new keyword list.
What happened
Workday’s October workforce report says mentions of basic AI skills in requisitions across more than 550 employers fell 25% after January 2026, while hands-on AI-building skills rose 51% from September 2025 to July 2026.
Why it matters
The figures suggest generic prompting is losing value as a differentiator, but they do not show which training causes better performance. Employers need observable evidence that separates literacy from reliable delivery.
Workday’s 5 October Global Workforce Report says demand for basic AI skills in requisitions across more than 550 employers peaked in January 2026 and then fell 25%. Mentions of building AI tools, automating workflows and AI engineering rose 51% between September 2025 and July 2026. An independent HR technology analysis describes the report’s mix of customer workforce data, requisitions and surveys, while noting that the dataset reflects Workday customers rather than the whole labour market.
Replace labels with an evidence ladder
Do not translate the finding into “prompt engineering is dead.” Requisition text is a demand signal, not a direct measure of skill quality, pay or job performance. Employers may also have stopped naming basic use because it is assumed, or because terminology changed. The data cannot distinguish those explanations.
Build an assessment ladder instead. At the first level, candidates should frame a task, check sources and identify unsafe inputs. At the second, they should automate a bounded workflow with tests, exceptions and a human handoff. At the third, they should monitor the workflow, diagnose drift and document who can change it. Score the work product, not the fluency of a tool demo.
Training portfolios need the same shift. Keep broad AI literacy, but attach it to real work samples: a reproducible analysis, an approval-aware automation and an incident review. Compare completion quality, correction burden and maintenance cost over time. The decision value is not in chasing the latest skill phrase; it is in making applied capability visible before hiring, promotion or redeployment decisions depend on it.