How the Index works

Collection

• A fixed sample of job ads is captured manually each month from SEEK, LinkedIn and Indeed.
• June 2026 baseline: 45 ads (SEEK 22, LinkedIn 17, Indeed 6), 44 unique advertisers.
• Editions are named for the hiring window they cover, not the capture date. Capture happens throughout the month, so this is the June 2026 edition.
• The sample is Melbourne-focused only.
• Each raw ad is preserved as captured, then converted into a structured research record (employer evidence, role metadata, exact quoted phrases) before any analysis.

What we count

• Advertiser mix by advertiser type, always with N.
• Role postures — evidence-based tags (technical builder, workflow automation generalist, analyst/bridge, adoption/enablement, data engineering, governance/leadership) assigned only from quoted ad text.
• AI-centrality — core / peripheral / title-only, because search keywords catch ads whose AI substance varies from "the whole job" to "one word in the title".
• Disclosed salaries — reported as a table of actual figures with N, not a median, until disclosure rates support one.
• Tools named — explicit mentions only, grouped into clusters (AI assistants, Microsoft automation stack, agentic/dev tooling, data platforms, RPA).

We designed a "role-conflation score" (how many distinct jobs one ad asks for, 1–6) before seeing any data. With only 4 SME-flagged ads this edition, publishing a scored distribution would have been statistically hollow, so stay tuned. Bundling is reported qualitatively with worked examples instead, and the score stays parked until the SME subsample reaches a usable size (indicatively 20+ ads).

Limitations (June 2026 edition)

• One month, one metro area, N=45. No trend claims before edition three.
• SME-flagged N=4 — worked examples only; no percentages or medians from that subsample.
• AI-title keyword collection undercounts SMEs, whose AI duties often sit inside operational titles rather than the job title.
• Employer sizes are inferred from public signals and may be wrong in individual cases; uncertain cases are excluded from the SME figures.
• Salary findings cover only the 7 of 45 ads that disclosed figures.

Cite this index

GraftPoint SME AI Hiring Index, June 2026 baseline edition. Fixed sample of Melbourne-area AI/automation job ads captured 3 July 2026, covering the June hiring window (N=45); 9% SME-flagged (under 200 staff, ABS definition). Employer-size rule v2, codebook v0.1. graftpoint.com.au/sme-ai-hiring-index/

Questions about the method, or want the underlying counts for a story? Email june@graftpoint.com.au.