A Working Framework — Companion to the AQ Maturity Model
Industry Benchmarks
Not every business needs AQ5. The right target — Destination AQ — depends on how repetitive the work is, how concentrated the expertise is, how tightly it's regulated, and whether the expertise itself is the product or just the delivery mechanism. Below is a sourced starting point for 21 industries, plus the gap calculator to find where your own organization sits against it.
How to read this: Destination AQ is the average of five drivers. Two are anchored to public data — Ops Leverage to McKinsey's automation-potential research, Regulatory Ceiling to the EU AI Act's high-risk categories — one (Bus-Factor Risk) is informed by BLS tenure data read alongside skill-intensity, and two (Decision Volume, Value Source) remain analyst judgment because no clean public dataset measures either. Every row below shows its actual sources — expand any industry to see them.
What actually anchors each driver
Data-anchored
Ops Leverage on Margin
McKinsey Global Institute, "A Future That Works" (2017) — technical automation potential by sector. <30%→1 · 30–39%→2 · 40–49%→3 · 50–59%→4 · ≥60%→5.
Data-anchored
Regulatory Ceiling
EU AI Act, Annex III — eight high-risk categories (biometrics, critical infrastructure, education access, employment, essential services, law enforcement, migration, judicial). Core activity listed → capped at 1–2.
Data-informed
Bus-Factor Risk
BLS median employee tenure by industry, Jan 2024. Read alongside skill-intensity — short tenure only signals risk when the role is expert-driven, not when it's commoditized.
Judgment
Decision Volume & Value Source
No public dataset cleanly measures either. Scored by analyst judgment, informed contextually by the same sources where relevant.
A case worth noticing: independent restaurants and boutique hospitality score the highest technical automation potential of any sector in the McKinsey data (73%) — yet Destination AQ still lands low, because Value Source pulls the average down. High automation potential alone doesn't mean full AQ is the right target — that's the point of averaging five drivers instead of chasing the biggest one.
Destination AQ by Industry
Vol Decision Volume
Bus Bus-Factor Risk
Reg Regulatory Ceiling
Ops Ops Leverage
Val Value Source
Dest AQ = average of the five, 1–5
| Industry |
Vol | Bus | Reg | Ops | Val |
Dest AQ |
Sources & Rationale by Industry
Expand any row for the exact figures and citations behind its score.
Full Source List
- McKinsey Global Institute — "A Future That Works: Automation, Employment, and Productivity"January 2017. Technical automation potential by industry and activity type. mckinsey.com
- EU AI Act — Annex III, High-Risk AI SystemsBinding classification of high-risk sectors and use cases under Article 6(2). artificialintelligenceact.eu
- U.S. Bureau of Labor Statistics — Employee Tenure in 2024, Table 5Median years of tenure by industry, January 2024 (released Sept 26, 2024). bls.gov
- U.S. Census Bureau — Business Trends and Outlook Survey (BTOS)Biweekly, sector-level AI adoption rates by NAICS code. census.gov
- U.S. Census Bureau — "Large Firms With at Least 20 Employees Biggest AI Users"America Counts story, May 26, 2026 — cross-sector AI-adoption validation figures. census.gov
Refresh cadence: Only the Census BTOS validation figures actually update every two weeks — that's the survey's own release schedule. The McKinsey automation-potential figures are a static 2017 report, the EU AI Act Annex III list changes only with legal amendment, and BLS tenure is collected every two years. Re-scoring the driver columns biweekly would mostly re-run unchanged numbers; the honest sync target is the BTOS validation column, not the underlying Destination AQ scores.