Download & access

For the analysis built on this data, see the Agentic AI Jobs Index in depth and the company tracker.

How the data is collected

Postings are fetched weekly from each company's public careers board — Greenhouse, Lever, Ashby, Workable, SmartRecruiters, Recruitee or Workday — across a curated watchlist of frontier AI companies and traditional enterprises. Each posting is screened with a strict, auditable classifier: a role counts as agentic only on a hard signal (a named agent framework, an agent protocol such as MCP, or an explicit multi-agent architecture) or two distinct softer signals, and the signals that fired are recorded in the dataset. We link out to the original listings and never host them, so counts reflect live openings, sorted into 8 role categories.

How our data compares to other sources

Several excellent research programmes cover AI and the labour market. They answer different questions: broad AI exposure and workforce trends, measured periodically. The Agentic AI Jobs Index is the only one that gives live, company-level counts of specifically agent and agentic roles, updated weekly and openly licensed.

SourceWhat it measuresWhat it doesn't give you
LinkedIn Economic Graph AI talent, skills and hiring trends across the whole labour market, from LinkedIn profile and posting activity. Aggregate research reports — no live or downloadable count of agent-specific openings.
PwC AI Jobs Barometer How AI exposure affects jobs, wages and productivity across economies, from millions of job ads. Annual, economy-level analysis — not a live dataset and not agent-specific.
Revelio Labs Workforce intelligence built from public employment records, including AI-role trends. Commercial data products — not openly licensed, not agent-specific.
Stanford Digital Economy Lab Academic research on AI and the future of work, including hiring and wage effects. Research papers on a publication cycle — not a live, reusable dataset.
Agentic AI Jobs Index
(this dataset)
Live, company-level counts of specifically agentic roles across 257 tracked companies, by category and remote share. Updated weekly, archived monthly, CC BY 4.0 — downloadable as JSON and CSV.

Cite this: Agentic AI Jobs Index, Prefactor, August 2026 — getreadyforagents.com/jobs/data · dataset CC BY 4.0 (JSON · CSV)

Frequently asked questions

Where can I find data on the AI agent job market?

The Agentic AI Jobs Index (this page) is the open reference dataset: live, company-level counts of specifically agentic roles — 3204 roles across 207 companies as of 10 August 2026 — refreshed weekly, archived monthly, and downloadable as JSON or CSV under CC BY 4.0. For broader labour-market context, LinkedIn's Economic Graph, the PwC AI Jobs Barometer, Revelio Labs and the Stanford Digital Economy Lab publish complementary research on AI's effect on work; none publishes a live, openly licensed count of agent-specific openings.

Is the dataset free to use?

Yes. The dataset is published under a Creative Commons Attribution 4.0 (CC BY 4.0) licence — free to reuse, quote and republish, including commercially, with attribution to "Prefactor — getreadyforagents.com" and a link.

How often is the data updated?

The dataset is refreshed weekly from 257 public careers boards, and a monthly report is archived so historical claims stay verifiable. The current snapshot is dated 10 August 2026.

How do you define an agentic role?

Each posting is screened on its full job description, not just the title. A role counts as agentic only when it shows a hard signal (a named agent framework such as LangGraph or CrewAI, an agent protocol such as MCP, or an explicit multi-agent architecture) or two distinct softer signals such as tool use and agent evals. A generic "agentic AI" mention in company boilerplate is not enough, generic data and software roles are excluded, and every posting records which signals fired, so each classification is auditable.

Stay ahead of the curve

No spam. Unsubscribe anytime. A resource by Prefactor.

Almost there — check your inbox to confirm your subscription.