AI-skill pages × business-school courses: one shared frame

When an AI skill sounds relevant to a business course — what does that similarity really tell us? So we asked something narrower: how closely do sampled AI-skill pages from one marketplace align with business-school courses?

That distinction matters. Value is measured through outcomes — demand, wages, quality, fees, employment. We ask something different: do the same job topics appear in both document sets?

Three inputs: occupation and task text, course descriptions, and public skillsmp.com pages. Curriculum collection was automated and verified from official university sources. AI-skill pages came through a separate acquisition workflow.

Then both sets meet on one fixed map. The method fits 361 job topics — textual factors, from occupation and task text — then scores every course and skill document against it.

An interpretability screen keeps the 239 topics with one dominant pattern and a non-filler top word. And a topic only counts as shared when both sets pass the same presence rule.

166 of the 239 screened job topics — 69.5% — met both thresholds, within these sampled documents.

It does not mean AI can perform that share of a course — or a job. Textual alignment is not evidence of capability, automation, replacement, adoption, causal impact, demand, price, or monetary value.

Breadth and depth answer different questions. Breadth counts how many topics appear in both document sets. Depth asks where course alignment is stronger. That ranking is descriptive — sample sizes differ, and we ran no statistical significance test.

For universities — curriculum review and independent tool testing. For skill builders — capability and demand validation. Method, sources and limitations are on the research companion.