Overview

FacultaStudio reads job advertisements and measures what the work actually asks for. Everything on this page counts postings — individual job adverts collected from public job boards, not people, hires or open roles. One advert is one posting, whether nobody applies or a thousand do.

It exists to answer a practical question: if you have an hour, what is it worth learning? The panels below read that off what employers are currently asking for.

Not included unless you switch them on: 328,800 postings from Fortune 500 career sites and 68,167 postings from small-employer job boards and 14,055 postings from US federal jobs, collected since 27 September 2026. The published figures stay on the panel they have always used.

Collected 2026-10-14, postings first seen 2023-03-01 to 2026-10-14.

321,586postings (n = 321,586)

117,836AI-related, of n = 321,586

230boards

33,068employers

Business-deep 25,378: AI in Business, Sales

The two large circles are the two depth tests a posting can pass; where they cross it passed both, and outside them it passed neither. Their areas are their counts, and the 44% of the engineering circle sitting inside the business circle is why this is a Venn and not four separate things. Each region is labelled with academic programmes, not jobs. Business programmes are placed where most of their postings sit, so the business majors land in business-deep on their own numbers, and Information Systems lands in the overlap because no single region holds most of it: 17% business-deep, 14% hybrid and 19% builder-deep. No business programme lands in builder-deep, so that region names the degrees its jobs — software, machine learning, cloud — draw on. The job families inside the circles are counted in them but not named. The small circles are the fields the four quadrants do not describe. How far a circle reaches in is the share of that field already at that depth, its area is its size, and it is outlined in the colour of whichever circle holds more of it, unless the two are within eight points — those are dashed, because a lead that small is a coin flip. Hybrid counts toward both. The small circles re-cut postings the large ones already count, and a posting that involves two fields is counted in both, so their areas are not meant to add up to anything. Read the provenance before the shapes. Both halves are provisional: the business programmes ride on a classifier awaiting human validation, and the job families outside business are a rule set nobody has yet coded a sample against.

Look inside one program

Each circle is a program, placed by how many job titles it shares with the others; open one to see its own quadrants, overlaps and skills, or see all programs as a table.

Four quadrants of the same labour market

Bar length is the number of postings in a region. The filled part is the share of that region which is AI-related. Both are the last 90 days, by real posting date.

Business, AI-thinthe established business majors, unchanged
74,908postings35.8%AI-relatedMost postings here: Information Systems 22,867 · Business Management 7,250 · Sales 6,468
Business-deep AIan AI-in-business major
25,378postings41.0%AI-relatedMost postings here: Information Systems 7,782 · Sales 6,108 · Business Management 4,243
Hybrid AIan information systems major, revised for AI
8,079postings75.3%AI-relatedMost postings here: Information Systems 6,225 · Business Management 930 · Sales 755
Builder-deep AIa computer-science AI degree
10,434postings66.0%AI-relatedMost postings here: Information Systems 8,769 · Business Management 357 · Sales 305

What the four quadrants are cut on. The capabilities a posting asks for, read from its description. A role naming three or more distinct engineering capabilities has builder depth; three or more business capabilities gives business depth. Both is Hybrid, neither is AI-thin. There are four because two yes-or-no tests cross, which is what makes them quadrants rather than a list. The threshold of three is the one the source analysis used, adopted rather than reinvented so the two stay comparable.

These 118,799 postings are the ones carrying a real posting date from the last 90 days outside the United States; the panel above is larger. Geography is a choice here rather than a fixed filter: it was pinned to the United States until 22 September 2026, which set aside about 44% of the scored frame without saying so. The window is deliberate: collection began in September 2026, so anything older is visible only if it stayed open long enough to be caught. Quadrants do not depend on the discipline classifier, so unlike the per-discipline figures they do not wait on human validation.

What each quadrant asks for

Ranked by how far Builder-deep AI sits from the other quadrants, not by how often the item appears. The bar is this region; the ticks are the others. An item everybody asks for tells you nothing about where you are.

Dispositions

O*NET work styles: how somebody works rather than what they know. A framework laid over the terms, not the terms themselves.

The clean comparison on this page. O*NET work styles take no part in the quadrant cut, so a gap here is a finding rather than arithmetic.

Dependabilityabout 9 in 100 postings ask for this
8.7%−13.8 points vs the other three
Stress Toleranceabout 12 in 100 postings ask for this
12%−10.3 points vs the other three
Social Orientationabout 3 in 100 postings ask for this
3.2%−9.4 points vs the other three
Analytical Thinkingabout 25 in 100 postings ask for this
25%−8.8 points vs the other three
Attention to Detailabout 10 in 100 postings ask for this
9.9%−7.5 points vs the other three
Innovationabout 24 in 100 postings ask for this
24%−6.3 points vs the other three
Initiativeabout 42 in 100 postings ask for this
42%−5.8 points vs the other three
Cooperationabout 64 in 100 postings ask for this
64%−4.7 points vs the other three

Shares are of postings in this region over the last 90 days, by real posting date, in the geography chosen above. These figures were fixed to the United States until 22 September 2026 and could not follow that choice; they can now.