Prepare Your Power BI Semantic Model for AI & Copilot

**Field descriptions (sample):** - `Net Revenue`: Total sales after discounts and returns, in USD. Excludes intercompany and contra-revenue items. - `Gross Margin %`: Gross profit divided by Net Revenue. Use this — not Revenue — to compare profitability across segments. - `Close Date`: Date a deal was marked Won in Salesforce. Use for period reporting; distinct from Invoice Date. - `ARR`: Annual Recurring Revenue. Annualised value of active subscriptions; excludes one-time fees. **AI Instructions text:** > "This model covers B2B Sales performance for the Americas region. 'Revenue' always means Net Revenue after returns. For 'targets', use the [Quota] table, not [Forecast]. Our fiscal year starts February 1 — when asked about 'this year' or 'YTD', use fiscal year. Default to current fiscal period unless the user specifies otherwise. 'Churn' refers to the [Churned ARR] measure." **Synonyms:** - Net Revenue → Sales, Bookings, Income, Top Line - Gross Margin % → Margin, Profitability, GM% - Churned ARR → Churn, Lost Revenue, Cancellations **Gap analysis:** - ✅ 'What is our revenue this quarter?' — answerable - ✅ 'Which region is furthest behind target?' — answerable - ❌ 'Why did churn spike in March?' — needs a narrative/commentary measure; model has no causal layer - ❌ 'What will Q4 look like?' — no forecast table connected; add [Forecast] table - ⚠️ 'Show me deals at risk' — partial: pipeline data exists but no risk-score measure