Documentation.

Categories become standards because they measure something the industry previously could not. NPS measured loyalty. OKRs measured alignment. The Autonomy Index™ measures workflow ownership — the depth to which AI has replaced or absorbed human work, rather than sat alongside it.

The framework is built to be applied consistently across industries, teams and workflow types, and to accumulate benchmark data over time.

The Autonomy Index™ scores four categories, weighted:

• Adoption — 20%
• Capability Enablement — 20%
• Workflow Ownership — 40%
• Strategic Impact — 20%

The composite produces a 0–100 Autonomy Score™ mapped to a maturity level from Manual to AI-Native Workflow.

A companion 2×2 that plots autonomy against output quality. Every workflow lands in one of four quadrants: AI Theatre, Useful Assistance, Risk at Scale, Trusted Workflow Autonomy.

Every workflow is decomposed into twelve stages: intake, triage, research, analysis, drafting, decisioning, execution, quality control, reporting, escalation, learning and optimisation. Each stage is scored independently from Manual to Fully AI-Owned.

Category scores are computed as the mean of their questions, normalised to 0–100, then combined using the weights above. Workflow Ownership is computed as the mean of the twelve stage scores.

Human Correction Load™ (used by the Autonomy Quality Matrix™) captures the proportion of AI output that still requires meaningful human rework.

• Workflow ownership — the share of stages executed end-to-end by AI without human intervention.
• Autonomy — how independently AI performs a stage.
• Quality — how safe, useful and commercially valuable the output is.
• Maturity level — the composite stage: Manual, Assisted, Augmented, Autonomous, AI-Native.

Benchmark ranges and 12- and 24-month target thresholds for each category and stage.

• v1.0 — Initial public release of the Autonomy Index™ and Autonomy Quality Matrix™.

A curated bibliography of adjacent work in AI adoption measurement, workflow analysis and organisational transformation will be published alongside the first annual State of AI Workflow Adoption Report.

Longitudinal follow-up assessments at 30, 90, 180 and 365 days will be used to validate the framework's predictive value and to establish industry benchmarks. Aggregated, anonymised data will inform the annual State of AI Workflow Adoption Report.