Context Gap - Governance Assessment

When different dashboards calculate "Revenue" or "Churn Rate" using different formulas or different source tables, leadership stops trusting the numbers. This recipe scans every report column across your dashboards, groups together fields that represent the same business concept, and flags exactly where the formula or the data source disagrees — then hands governance teams a ready-to-work action list instead of a vague warning.

What you receive

How the recipe works (guided flow)

Step 1 — Reporting Surface Inventory
Pulls every active report column from all charts and dashboards, along with its formula and ownership details, into one master list.

Step 2 — Normalize Report Column Name
Cleans up naming differences (like "Day of Month" vs "day_of_month") so variations of the same metric name are grouped together.

Step 3 — Attach Formula and Source Signature
Traces each report column back to its calculation formula and the exact source table and column it pulls from.

Step 4 — Profile Definitions by Normalized Name
For each grouped metric name, counts how many different formulas and different source mappings are in use across reports.

Step 5 — Detect Definition Conflicts
Flags metric names that have more than one formula, more than one source, or both — separating genuine conflicts from consistent, stable metrics.

Step 6 — Build Conflict Summary
Creates a condensed list of only the conflicted metrics, ranked so the most inconsistent ones surface first.

Step 7 — Build Conflict Detail Cases
Expands each conflict into the individual report columns, formulas, and sources that are causing the mismatch, so teams can see the exact evidence.

Step 8 — Add Governance Context
Attaches the owner, steward, and custodian for each conflicted report column, plus the report name and type, so it's clear who should act.

Step 9 — Stable Metric Definitions
Separately identifies metrics that are already consistent everywhere — a trusted shortlist worth promoting into the official business glossary.

Step 10 — Metric Definition Prioritization View
Ranks every metric — conflicted or stable — by how severe and how widely-used the issue is, so the riskiest ones rise to the top.

Step 11 — Metric Definition Action Queue
Builds the steward-facing to-do list: each conflicted metric gets a priority rating (Critical/High/Medium/Low) and a plain-language recommended fix.

Step 12 — Display Result
Shows the full action queue sorted with the highest-priority issues at the top, ready for review.

Step 13 — Create unique metrics aggregate table
Consolidates all the detail into one row per metric, combining every report, owner, formula, and source variant into a single clean summary.

Step 14 — AI Decides Metrics Validity
Uses AI to judge whether each grouped name is actually a meaningful business term (like "revenue" or "account_id") versus meaningless noise (like a stray number or code), so junk doesn't clutter the results.

Step 15 — Filter valid terms
Keeps only the AI-approved, business-meaningful metrics for the final analysis, discarding noise automatically.

Step 16 — Metrics with highest Impact and Formula Conflicts
Surfaces the top 10 metrics with the most source inconsistencies, visualized as a bar chart comparing formula count against how widely each metric is used.

Step 17 — Review Priority Distribution
Shows a pie chart of how many metrics fall into each priority level (Critical, High, Medium), giving leadership an at-a-glance view of the overall governance workload.

Sample insights (illustrative)

Make sure the following ingredients are available in your workspace:

• Connected datasets are crawled and profiled
• Columns have distinct counts and top values generated