---
name: powerbi-anomaly-to-explanation
description: Use when Power BI must explain anomalies, outliers, sudden KPI changes, margin drops, cost spikes, demand shifts, inventory changes, production losses, or cash deviations.
---

# Power BI Anomaly to Explanation

Use this skill after an anomaly is detected or suspected.

## Explanation Pattern

- Detect: metric, period, entity, threshold, baseline.
- Segment: product, customer, supplier, plant, channel, account, region, owner.
- Decompose: mix, price, volume, cost, timing, currency, one-time event.
- Validate: source totals, postings, data quality, late records.
- Explain: top drivers and confidence.

## Output Requirements

- Include anomaly definition, baseline, driver contribution, confidence, and validation status.
- Flag data-quality anomalies separately from business anomalies.