Schema creates structure. Semantics creates meaning. Brando operationalizes both.

The distinction

Schema defines the data model. Semantics defines the intent behind the model.

An organisation can have a perfectly tidy schema and still fail operationally if nobody agrees what the fields mean, how the rules should be interpreted, or what behaviour the system is supposed to support. Structure alone creates order. It does not create understanding.

That is why both layers matter in AI governance:

• schema tells the system how information is organised
• semantics tells the system what the organisation means by that information

Why this becomes a business problem

In most enterprises, the gap between schema and semantics is hidden by human judgement. Experienced people know what the policy “really means”. They know when a phrase is acceptable, when a claim needs escalation, and when a context switch changes the right answer.

AI systems do not have that implicit organisational memory. If semantics are missing, they may execute something structurally valid but operationally wrong.

How the Brando uses both

Within the Brando, schema supports the structured layer: datasets, records, field definitions, policy objects, and machine-readable specifications. Semantics supports the governed layer: intent, meaning, context, and interpretation. The operating model matters because it keeps those two aligned over time rather than letting them drift apart.

In practical terms, that is what allows an organisation to:

• define policy in a form systems can use
• preserve the meaning behind those definitions
• test whether implementation still matches intent
• revise safely when the business changes