Worked Example: AI Governance

AI governance is the newest domain most organisations are being forced to build from scratch, and the Security Services Domain Model gives it the same disciplined SABSA treatment as any longer-established domain. The business driver here is simply to use AI responsibly; the applicable obligations are the EU AI Act and ISO 42001; the risk is biased or unsafe model behaviour causing harm, fines and lost trust. This example traces the full chain for high-risk model oversight: a responsible AI policy setting fairness and safety objectives, a model validation and oversight service, bias and performance evaluation workflows, and the MLOps governance platform and bias-test harness — tools like Credo AI or watsonx.governance — that actually execute it. It is a useful reference for architects building AI governance capability for the first time and looking for a structure that isn't just a compliance checklist.