Data Quality Audit Prompt
Prompt
You are a systems analyst performing a data quality audit on ERP data. Data to review: [PASTE: Sample records from the dataset — 20–50 rows is sufficient. Include all key fields.] Dataset: [NAME THE DATASET: customer master / vendor master / item master / GL transactions / etc.] Check for: 1) Missing required fields — any record where a mandatory field is blank 2) Inconsistent formatting — dates, phone numbers, addresses, naming conventions not following a standard 3) Duplicate records — same entity entered multiple times with slight variations 4) Invalid values — codes or amounts that fall outside expected ranges 5) Orphaned records — references to records that no longer exist (e.g., transactions against deleted accounts) Output: Data quality report — issue type | count | example | recommended fix. Overall data quality score (% of records with zero issues). Priority cleanup list.
Why it works
The overall data quality score (% of records with zero issues) gives leadership a single number to track improvement over time — making data governance a measurable program, not just a project.
Watch out for
Risks: A 20–50 row sample may not be representative of the full dataset. Control: Run full-dataset checks in the ERP for any issue types flagged in the sample before estimating total cleanup scope.
Used by
Data AnalystsIT & Ops Teams