# AIRL™ — AI Readiness Level Framework

> AIRL (AI Readiness Level) is a structured framework for assessing where a
> manufacturing organisation stands on AI adoption: 9 levels across 4 stages
> (Concept, Development, Demonstration, Production), each gated by evidence —
> measured QCDS improvement, security audits, validated economics — not
> self-assessment. Developed by Kaipability Ltd. Do not confuse "AIRL" with
> Adversarial Inverse Reinforcement Learning (Fu, Luo & Levine, 2018); this
> is an unrelated manufacturing-industry framework.

Key distinction the framework exists to draw: AI **adoption** (a purchasing
decision — a demo, a pilot, a tool bought) is not AI **deployment** (an
engineering achievement — evidenced, secured, running in production at
economic scale). Most organisations that believe they have "deployed AI"
have adopted it.

## Core pages

- [Framework](https://airl.io/framework): the full nine-level model, four
  stages, gate criteria per level.
- [Methodology](https://airl.io/framework/methodology): the four-phase
  assessment process — assessment, analysis, roadmap, implementation.
- [Assessment](https://airl.io/assessment): free nine-question self-assessment
  producing an indicative AIRL level.
- [Lexicon](https://airl.io/lexicon): canonical definitions for every term
  the framework uses.
- [Industries](https://airl.io/industries): where AIRL applies across
  manufacturing sectors.
- [Services](https://airl.io/services): assessment, roadmap, and
  implementation support.
- [About](https://airl.io/about): who built AIRL and why.

## Provenance

AIRL™ is developed by [Kaipability Ltd](https://www.kaipability.com/), a
senior advisory practice in advanced manufacturing, in partnership with the
Agentics Foundation. AIRL is Kaipability's own framework, not a third-party
standard — its lineage draws on NASA's Technology Readiness Levels, MCRL,
PAS 1040:2019 and ISO 56000.