Add business glossary as first-class graph entities

[ravisuhag]

Context

Technical metadata alone doesn't tell AI agents what things mean in business terms. A business glossary links business concepts (revenue, churn, LTV) to the technical entities that implement them.

Scope

• New entity type for business terms/concepts
• Each term has: name, definition, domain/category, owners, synonyms
• Relationships: business term → implementing entities (tables, columns, dashboards, metrics)
• Search should surface business terms alongside technical entities
• Context composition should include business term definitions when relevant

Examples

• "Revenue" → linked to monthly_mrr table, revenue_dashboard, and the net_revenue column in orders
• "Churn" → linked to the agreed definition by the analytics team, the SQL that calculates it, and the dashboard that displays it

Why

When an AI agent needs to reason about "revenue," it should find both the business definition and every technical implementation. The glossary bridges business language and technical metadata.

References

• Compass Roadmap — Knowledge Layer