FIELD RELAY / CASE STUDY

A review desk for the difference between two exports.

Field Relay is a browser-local reconciliation desk for comparing a baseline export with a candidate export before someone acts. It keeps exact input, original source rows, field mapping, and local review decisions in an explicit download.

Comparison design

Each export has its own explicit business-ID map and selected comparable fields. IDs remain literal strings, so leading zeroes are preserved. One-to-one records can be unchanged or changed; records only on one side stay visible. Duplicate or blank IDs stay ambiguous with all row evidence and never receive an automatic replacement decision.

Reviewable evidence

A reviewer can record an explicit hold, a proposal to use Candidate as current, or an expected difference, with a reason. These decisions bind to the exact local input and mapping revision. Editing either export, a map, a compared field, or whitespace normalization clears them. The report includes full input text once in a shared comparison reference, then maps review decisions to that reference with mappings, findings, and unresolved items. Diagnostic fingerprints are labels only.

Separate delivery lab

The earlier fictional delivery exercise remains separate from reconciliation. It demonstrates parsing, approval, outage retry, and a replay-safe ledger, but it does not consume comparison results or update any downstream system.

Operational boundary

The reconciliation desk reads pasted or browser-selected files only in the active tab and never sends them to a server. Exports are explicit downloads. The included examples and delivery lab are fictional; the comparison workflow works with a user’s own CSV exports.

Open the reconciliation desk · Interview walkthrough · Inspect the source