Domain & Semantic

• Domain ontology and taxonomy — entities, relationships, business rules
• Semantic layer and metric definitions so one term returns one number
• Glossary and synonym mapping between business and system vocabulary
• Regulation as knowledge: rules the agent must understand, distinct from policy it must obey
• Stewardship: who owns each definition and how changes are reviewed

Accountable owner

Data Governance / Enterprise Architecture, with domain SMEs

• Chief Data Officer or Head of Data Governance — accountable owner
• Ontologist / taxonomist — models the domain
• Data steward per domain — owns definitions and approves changes
• Domain subject-matter expert — supplies the ground truth
• Semantic engineer — implements and tests the layer

Ownership and phase are architectural recommendations, not assessed results.

Start with the twenty terms that appear in disputed answers, not a full enterprise ontology. Expand only where ambiguity has actually cost something.

Signals to watch

• Coverage of disputed terms with an approved definition
• Definition change rate and steward review time
• Metric disagreement incidents per quarter
• Share of agent answers traceable to a governed definition

Glossary and ontology changes need domain-owner review. Preserve semantic compatibility and propagate changed definitions explicitly.

Failure modes

1. Semantic drift: definitions diverge quietly between teams and the agent answers confidently from the wrong one
2. An ontology nobody owns after the consultants leave
3. Regulation held only as prose, so nothing downstream can act on it
4. Modelling the whole enterprise before proving value on one domain

Existing tool and standards snapshot · approximately mid-2026. Examples, not required dependencies; verify current versions before adoption.

This harness is what stops a competent-sounding agent from being subtly wrong about your business. It is also the slowest to build, which is why it belongs in phase 2 rather than the pilot.