Overview

AI9GM (ai9gm.com) is the AI-Centric IT Governance Model. It is an open reference model for governing AI across the estate an AI system depends on. It organizes governance into six interdependent layers (Foundation, Structural, Intelligence, Control, Execution and Strategic). It extends established standards such as COBIT, ITIL and ISO into the AI era. It does not replace them.

Decision rights set the published model apart. AI9GM names 94 decisions with permanent identifiers. For each decision it states who Decides, who is Consulted, who Executes and what Evidence must be retained. Standards often leave that allocation implicit. AI9GM makes it the primary reading surface. Layer 4 (Control) holds a large share of the catalog. Every layer still carries its own rights.

AI9GM and StrataProtocol.org were developed in parallel and informed each other. An early shared bet was a Docus-style documentation portal. StrataProtocol.org shipped on Nuxt and Docus. The AI9GM publication site later shipped on Astro 7 (Lotus theme) under the same STRATA delivery discipline. The parallel is methodological. It is not a claim that both products share one frontend stack today.

Classification: Class 1.3 Platformization

AI9GM is classified Class 1, Category 1.3 Platformization, the same category as StrataProtocol.org. The work did not invent an unproven product. It took a governance model refined through practice and converted it into an accessible, externally facing standard.

That is the Platformization move: a validated capability becomes a public, reusable platform. Here a governance model becomes a versioned, citable framework with permanent addresses for its decisions.

Technology Stack

STRATA Principles Observed

AI9GM is unusual among these cases. Its subject is layered governance, so the parallel with STRATA is structural, not only methodological.

• A layered governance model. AI9GM's six interdependent layers mirror STRATA's stratum model. Governance fails when its concerns are tangled. It holds when they are layered so each builds on the one beneath it.
• Decision rights as the binding surface. Permanent identifiers and an explicit Decides / Consulted / Executes / Evidence allocation turn a layer diagram into something an organization can operate and correct against.
• Self-documenting by design. The governance framework for AI is published under STRATA to a high documentary standard. The live site is both the model and the record of how it is structured.
• Honest maturity. The published site states what v0.9 is and what it has not earned (including certification and community governance). That refusal is part of the proof.

Delivery velocity under STRATA

The AI9GM publication MVP ran under STRATA from 22 August 2026. In five weeks the repository took 117 commits. That count includes engineering work and the operational work the release required: DOI registration for the archival record and Coolify deployment of the Node host. The point is not the commit total as vanity. It is that a governed AI-assisted loop carried a Class 1 site from adopted specification through permanence, write path, discoverability and production acceptance without inventing a second methodology mid-flight. STRATA Protocol works when the artifact trail and the human gate stay intact.

What STRATA Would Have Formalized

AI9GM was self-documenting by design. Much of its early construction discipline lived in practice rather than in a named protocol. STRATA makes that explicit for later work: a written Classification Record fixing the platformization stance, a derivation chain from strategic vision down through the governance layers and a five-stratum artifact trail for how the framework is published, not only what it prescribes. The model already taught governance. STRATA now governs the publication delivery the same way this platform does. See the Execution Loop.