01
Column mapping
Map uploaded PAS fields onto the target calculation schema.
New extracts can be onboarded without rewriting the engine.
The brief
Why it exists, and what it changed.
The problem
Death, maturity, and surrender calculations lived in spreadsheets that could not be versioned, tested, or explained at policy scale.
What we built
Build an ops cycle — upload, mapping, data quality, run, results, exceptions — on a Fastify engine with Postgres-backed tables.
What changed
Runs produce auditable outputs, exception queues, and policy illustration PDFs from a single active configuration.
01
Spreadsheet calculation risk
Formula and factor logic was copied across workbooks with no shared version history.
Impact · Errors only appeared after a run had already been used downstream.
02
Extract chaos
PAS files arrived with shifting columns, encodings, and incomplete rows.
Impact · Analysts spent more time cleaning files than reviewing results.
03
Exception handling
Failed policies had no owned queue, severity, or re-run path.
Impact · Reconciliation became a side conversation instead of a product workflow.
