From the first critical reflex to governed-agent architecture.
The proprietary framework for qualifying AI use at work. Six belts, three reading zones, a foundation aligned with the AI Act literacy obligation. We don't certify an impression: we certify proof.
Recalibrated in 2026, the framework shifted its centre of gravity. Prompting is no longer a level but a prerequisite; the foundation becomes the critical reading of outputs and data security; the top qualifies agent supervision and governance. The "Prompt" suffix is gone: progression reads as belts, like in Lean, but AI-native.
Literacy and security. The legal floor required by Article 4 of the AI Act.
Value applied to the business, delegation judgement, reusable assets.
Governed systems, agent supervision, compliance. The agentic turn.
They aren't a level: they run through every belt and deepen with each zone.
Every level is validated on three pillars. No output counts as validation without having been judged.
Since February 2025, every AI deployer must ensure a sufficient level of AI literacy among its staff; supervision by national authorities begins in August 2026. The framework's Foundation matches this obligation exactly. ANGLE 360™ certification produces the documented proof expected (level, evidence, date, progression trace) and turns a regulatory constraint into a readable progression system.
A mature organisation builds a broad White and Yellow base, a Green and Blue operational core, a few Black referents, and a very limited number of Gold architects. The higher the belt, the more it is proven by real work and governance, and the less by theory.
Industry, support functions, management, projects, HR, QHSE, procurement, marketing, transformation. The standard deploys at the scale of a site, a department or a group, with shorter validity periods at the top of the scale: 24 months at the Foundation, 12 months in Architecture, where agentic practice moves fastest.
