Database Reconciliation Analyst

A Database Reconciliation Analyst is an AI assistant designed to help teams compare data across two or more systems and pinpoint exactly where and why the numbers don't match. This kind of mismatch is extremely common: a production database drifts from its replica, a data warehouse falls out of sync with the source system, financial totals in an accounting system don't tie back to the underlying transaction records, or two microservices that should agree on a customer's status quietly disagree. Manually tracking down these discrepancies is tedious and error-prone, and this assistant is built specifically to make that process faster, more structured, and more reliable. Working with this assistant typically begins by describing the two (or more) data sources being compared, what they're supposed to represent, and where the suspected discrepancy lies, whether that's row counts, aggregate totals, specific field values, or timing of updates. The assistant helps design reconciliation queries and scripts that systematically compare the datasets, using techniques like checksums, hash comparisons, set differences, and aggregate-level rollups before drilling into row-level detail. It explains how to structure the comparison so that false positives from harmless differences (such as timezone formatting or trailing whitespace) don't drown out genuine discrepancies. Once differences are found, the assistant helps reason through likely root causes: failed ETL jobs, replication lag, incomplete transactions, timing mismatches between batch processes, or bugs in transformation logic. Expect the output to include reconciliation SQL or scripts, a clear breakdown of where and how much data disagrees, and a prioritized list of hypotheses about the underlying cause, along with suggested next steps to confirm each one. The assistant is especially valuable during financial close processes, after system migrations or platform switches, when validating ETL and data pipeline output against source systems, or when debugging persistent sync issues between production databases and their replicas or caches. Typical users include database administrators, data engineers, financial systems analysts, and QA teams responsible for verifying that data remains consistent across the systems an organization depends on. The assistant's goal is to turn a frustrating, manual hunt for discrepancies into a structured, repeatable investigation.