Data Quality Scoring Reconciliation Assistant
Description
Match, explain, and prioritize differences for data quality scoring; draft root-cause narratives and recommended correcting actions for analyst review. The MVP would connect Advana, data catalog, ERPs, feeder systems, data dictionaries, lineage tools and produce read-only recommendations for CDAO, OUSD(C), Component Chief Data Officers, system owners.
AI / analytics pattern
rules + ML matching + GenAI explanation
Automation level / stage
assisted reconciliation
Expected benefit
Reduced manual reconciliation effort, faster close, stronger audit trail for differences.
Audit / financial statement impact
Audit traceability and all statements dependent on source data quality
Controls / human review
Human review for exceptions and recommendations; maintain evidence packages, lineage, source citations, model cards, data-quality checks, and periodic QA sampling.
Data needed
Advana, data catalog, ERPs, feeder systems, data dictionaries, lineage tools; master/reference data; audit logs; policy/control requirements; prior exceptions; relevant document evidence.
Possible metrics
match rate; aged differences cleared; analyst hours saved; unsupported adjustment reduction
MVP scope
Start with one Component/reporting entity and one subprocess (data quality scoring) for two close/audit cycles; read-only outputs first.
Related material weakness / control objective
Universe of Transactions, data quality and system modernization