The AI Job Atlas
April 2026 · by Alex Ilin
Your next job posting probably mentions AI. But which tools? At what level?
We read 140,000 listings to find out.
One email a month, when the next edition drops. Occasional targeted insights between.
Each company counts once per role, so no single employer dominates the numbers.
We show all companies together — but flag the AI-native ones
separately, so you can see where the frontier is and how far the mainstream has come.
ChatGPT
Copilot
"High level of comfort using AI and automation tools (e.g., ChatGPT, copilots, workflow automation)."
The world's largest employer now lists AI fluency as a basic EA skill.
AI rises with seniority. Almost everywhere.
Thirteen of sixteen roles ask for AI more at the senior level than at the junior one.
The strongest climbs aren't in engineering — they're in Product and Design.
Early/Mid
Senior+ (arrow points to senior)
Δ in percentage points
Each company counts once per seniority band. Companies where AI appears in every listing
regardless of role are excluded. Seniority is inferred from the job title. Rows with fewer
than 80 companies in either band are hidden.
AI requirements, role by role.
For each role, the bar shows what share of companies mention AI in their listings —
all companies, not just tech startups. The ● dot
marks the AI requirement rate at AI-native companies — the upper bound for AI adoption
in that role. Where the dot leads the bar, AI-native companies are ahead of the mainstream.
Tech roles
Non-tech roles
AI-native companies
% of companies that mention AI tools in at least one listing for this role. Each company
counts once. Bar = all companies. Dot = AI requirement rate at AI-native companies,
after filtering out company-wide AI language. Only shown for roles with 10+ AI-native employers.
Claude
ChatGPT
"Manage and improve AI-assisted contract workflows (Claude, ChatGPT)."
Named tools, specific use case. Not an AI company — a SaaS company whose legal team adopted AI for contract work.
Startups vs Enterprises
How much of AI adoption is a startup phenomenon? We compare startups and tech
companies against large enterprises — hospitals, banks, manufacturers.
Startups = companies on Greenhouse, Lever, Ashby. Enterprises = Workday companies. Showing roles with the largest gap. Average row is company-count-weighted across all classified roles.
ChatGPT
Copilot
"Experience using AI tools, including generative technologies (e.g., ChatGPT, Microsoft Copilot)."
AI has reached the hiring desk at Big Pharma — a senior role at a traditional employer.
Copilot
AI tools
"Develop working proficiency in audit and AI tools (e.g. Workiva, Copilot)."
Copilot listed next to Workiva — AI as mundane infrastructure, not a special skill.
Next month's edition, in your inbox.
We're updating this monthly, with targeted insights between editions when we spot something worth sharing.
Subscribe to get the next one when it drops.
Methodology
- Corpus: ~140,000 active job listings scraped from ATS feeds (Greenhouse, Lever, Ashby, Workday) of ~3,000 companies, April 2026.
- Counting unit: Unique companies per role, not listings. One company = one vote per role, regardless of how many listings they post.
- Role classification: Regex on job titles into 26 clusters. Unclassified titles (~68%) excluded.
- AI detection: Pattern matching on description text for AI tools, frameworks, and task language. 35 named tools/platforms detected by regex. Suppressed when fewer than 3 mentions.
- Company-wide AI language: For companies with 10+ listings, if AI text appears in >70% of listings, that language describes the company — not the role. These companies are flagged as "AI-native" and their genuine per-role AI requirements are measured separately. Validated at 96% accuracy.
- Analysis: For each role, we measure what share of companies mention AI tools in at least one listing. We compare this across seniority bands (Early/Mid vs Senior+), company types (startup vs enterprise), and between AI-native and mainstream employers. The AI-native rate serves as the upper bound for where each role is heading.
What we're watching next
This is the April 2026 baseline. We'll update monthly. Three things to track:
- Is AI tool adoption growing in non-tech roles — or has the adoption line stalled?
- Agentic AI shows the smallest gap between tech and non-tech adoption. Is it closing further?
- Healthcare is the holdout — Nurse at 5%, Therapist at 0%. Does that change?
by Alex Ilin · All labs