AI Architecture Practice · Australia
The AI estate of your organisation,
structured with architecture discipline.
Kettera treats artificial intelligence as an estate to be mapped and governed — not a pile of pilots. Enterprise-architecture method supplies the rigour; AI is the entire subject; the Australian regulatory context is the frame.
What is the current and future state of AI in this organisation, and how do you move from one to the other defensibly?
The front door
Tools, not articles, are the reason to arrive.
This audience does not browse EA websites. Two instruments do the work: a deterministic, transparent assessment of where your AI estate stands, and a classifier that tells you which AU obligations a given use case trips.
AI EA Maturity Assessment
Score your estate across six axes — five domains plus Governance. Radar, an overall Kettera Index, three prioritised gaps, and routing into the matched content. Deterministic scoring, in your browser.
Run the assessment →AI Use-Case Risk Classifier
A short flow maps a use case to its EU AI Act tier, whether it trips the DTA impact-assessment threshold and the Privacy Act ADM position — then lists the obligations that attach. Deterministic, and runs in your browser.
Open the classifier →How the practice is organised
A matrix, not a reading list.
Two axes. Method is the journey — Baseline, Ambition, Strategy, Roadmap, Govern, Operate. Domains are the pillars you look at, each treated purely as an AI lens. Every cell is a place where work actually happens.
What we go deep on first
Six anchor artefacts.
One per domain, each built to authority standard and Australian-contextualised — the precise things an AU ICT head needs this quarter.
AI app portfolio inventory
Ties to the DTA AI use-case register obligation.
Data readiness audit
The precondition nobody scopes.
AI ambition canvas
Board-level posture tool.
Operating-model transition
CoE vs federated, decision rights, accountable officials.
Platform roadmap
Sovereign vs hyperscaler, compute, exit / portability.
AI governance operating model
Review boards, gates, risk classification, assurance — the hub the Risk Classifier links into.
The citation anchor
One canonical Australian AI regulatory reference, kept current.
AI6, VAISS, the DTA mandatory obligations, the Privacy Act ADM changes, ISO 42001, the EU AI Act — structured, dated, and crosswalked. It is a living page: it moves only when the regulation moves.
Current as at 25 June 2026