Expected 2,475
Reproducible synthetic test
5,000 rows.
Hard boundary.
We tested the same reconciliation workflow offered in the fixed pilot at its full combined-row limit, then tested the rejection boundary and an identical rerun.
All records were constructed for this test. The audit does not claim customer outcomes, financial impact or production-system performance.
01 / RESULTS
Expected counts.
Observed exactly.
2,500 source rows + 2,500 CRM rows. Shared identifiers were used under the same narrow rules as the pilot.
Expected 25
Expected 25
Expected 0
02 / CONTROLS
The limit is enforced.
The rerun is stable.
5,001 rows rejected
The workflow stopped before reconciliation and returned the agreed 5,000 combined-row scope limit.
Exact rerun reused
An identical input produced the same deterministic run identifier instead of creating an ambiguous duplicate result.
13 checks passed
The existing automated checks passed before the capacity run. This supports the tested rules; it does not replace review of client-specific definitions.
03 / WHAT THIS ESTABLISHES
A bounded workflow.
With visible limits.
- The tool can process the full row allowance in a controlled synthetic fixture.
- Expected matched and unmatched counts are preserved.
- Out-of-scope volume is rejected rather than silently accepted.
- Identical inputs can be recognized consistently.
It does not establish that two real exports are comparable. Join keys, field definitions, dates, time zones and source coverage still require scope review.
FIXED PILOT · $500 USD
Test the structure.
Before sending records.
Send the two systems, approximate combined row count, shared identifier and decision the review should support. Field names or synthetic examples are enough for the first check.
Request a scope check