Duplicate scan

AI duplicate scan for US CPAs and accounting professionals

Catch duplicate invoices, payments, and transactions before they hit the file. Each candidate comes with a confidence level and the source rows behind the match, so review is fast.

How it works

1. Step 1

   Upload the export

   Attach an AP, payment, or transaction listing as CSV or XLSX.

2. Step 2

   Matching runs deterministically

   Rows are grouped by amount, date proximity, vendor, and reference similarity.

3. Step 3

   Work the duplicate groups

   Review candidate groups by confidence, with the exact source rows behind each match.

What you get

• Duplicate groups
• Confidence
• Match basis

Frequently asked questions

Will it catch near-duplicates?

Yes — matches don't require identical rows. Same amount with a shifted date, a re-keyed invoice number, or a slightly different vendor name are all surfaced with lower confidence for review.

Is the output generated by an LLM?

No. The workpaper numbers are computed by deterministic spreadsheet logic — the same inputs always produce the same outputs. AI is only used to route your request and explain the results, so every figure is traceable to source rows.

What happens to my uploaded file?

Files are processed to build your workpaper and stay attached to your private conversation. They are not used to train models.