ABM Invoice Analytics: find the bills you have already paid.

Download

Pick your computer. Each download is the whole app, and it needs no internet connection once installed.

All versions and checksums: release page. Source code: github.com/abmodi-ai/ABM-Invoice-Analytics-Desktop-App.

What it checks

Every file you bring in is compared with everything already billed, and the app answers one question for it straight away: has any of this been billed before?

Duplicate invoices. The same invoice sent twice, re-keyed with a changed character, re-issued under a new number, split into parts, or the same file imported a second time. This covers bills you receive and bills you send.

Charges billed twice for the same patient. The same service on the same date on another invoice and, once Medicare reference data is imported, units above Medicare limits, procedures that should have been billed together, and repeats inside a follow-up period.

Clinical-trial site invoices. A protocol visit billed again on another invoice, a high-cost procedure repeated within days, two different visits dated the same day, and the same charge appearing twice on one invoice. It reads the site invoices themselves: subject IDs, visits, milestones and pass-through costs.

It reads PDF invoices, digital or scanned, CSV and Excel exports, and X12 837 and 835 claim files.

It flags; you decide. Nothing is paid, rejected or changed without a person.

How it works

It shows its working. Each flag names the rule that raised it and puts the two records side by side, with the matching and differing fields marked. It also lists the checks that could have explained it away, such as a credit note or a corrected claim. A reviewer confirms or dismisses it, and a dismissed pair is not raised again.

Your data stays on your computer. There is no account and no cloud service. The database is encrypted on your disk, with the key held by your operating system: Windows, the macOS Keychain or the Linux keyring. The app makes no internet connections.

The checks are ordinary code, not AI. Detection uses published rules you can read. Local AI is optional and switched off by default. When it is on, it only explains flags and suggests a verdict, running on your own machine, and its explanations are checked against the evidence before you see them.

Install

Windows

1. Run ABM-Invoice-Analytics-Setup-x64.exe.
2. Until code signing is complete, Windows shows Windows protected your PC. Click More info, then Run anyway.
3. Open ABM Invoice Analytics from the Start menu.

Mac

1. Open the .dmg and drag ABM Invoice Analytics onto Applications.
2. Until Apple notarization is complete, macOS says it cannot verify the app the first time. Click Done, open System Settings → Privacy & Security, and click Open Anyway.

Linux

1. Make the file executable and run it: chmod +x ABM-Invoice-Analytics-Linux-x86_64.AppImage && ./ABM-Invoice-Analytics-Linux-x86_64.AppImage
2. Ubuntu 24.04 and newer need sudo apt install libfuse2t64 once.

Then open Ingest, drop in your invoices, and review what it finds in the Review queue. Before you use it with real patient data, turn on Settings → Sign-in.

What it is not

It is not a payment system and does not stop a payment by itself. It is not a validated GxP system out of the box, and it does not replace your coders, auditors or quality unit. It finds likely duplicates and shows you why, so the people accountable for the money can decide faster and with the evidence in front of them.