Second Brain

Adopt phased Tableau ↔ Cortex Analyst alignment: (Phase 0-1) Definitional Contract — golden KPI parity test with Semantic View as canonical definition, Sales first then CRM, this quarter; (Phase 2) Ge

2 min read

Adopt phased Tableau ↔ Cortex Analyst alignment: (Phase 0-1) Definitional Contract — golden KPI parity test with Semantic View as canonical definition, Sales first then CRM, this quarter; (Phase 2) Ge

Decision

Adopt phased Tableau ↔ Cortex Analyst alignment: (Phase 0-1) Definitional Contract — golden KPI parity test with Semantic View as canonical definition, Sales first then CRM, this quarter; (Phase 2) Generated Reporting Views — RNWL_PROD.RNWL_REPORTING.VW_* regenerated from Semantic View metric/dimension definitions, target Q1 FY27; (Phase 3) Governance — release checklist + nightly parity job + KPI definition register; (Phase 4) Native Tableau semantic-view consumption piloted late FY27 when VizQL Data Service matures.

Rationale

Discovery confirmed two parallel reporting layers in Snowflake: Tableau queries 28 hand-built views in RNWL_PROD.RNWL_REPORTING (one per dashboard); Cortex Analysts (RUNWAL_BOT, RUNWAL_CRM_BOT) query Semantic Views V2 in RNWL_PROD.RNWL_SALES/RNWL_CRM. Both layers transform same upstream raw data independently — same measure (e.g. MTD bookings) computed two different ways with no enforcement of equivalence. Drift is structural, not accidental. Phased approach delivers parity in week 1 (validation), unification in Q1 FY27 (single source of truth aligned with CTO’s data-foundation focus shift), and prepares for FY27 H2 native consumption when Tableau platform catches up. Sales selected first because RUNWAL_BOT has been live since Nov-2025 (more mature than the 06-May-2026 CRM agent). Construction starts aligned by design — RNWL_CONSTRUCTION_REPORTING semantic view already exists; dashboards will consume it from day one when Bandhoo/Colab data arrives. 0.75

Alternatives Rejected

Outcome

Pending

Graph View

Backlinks