Accounts payable · Production

ACM Services automates AP invoice processing with Nanonets, eliminating manual batch data entry

The problem

ACM Services' AP team manually keyed invoice data from vendor batches of 1-300 invoices received three times per week, consuming 15-20 hours per batch and totaling 1,040 man-hours (~4 months) of manual data entry per year, leading to payment delays and an inability to take vendor discounts.

Workflow diagram · grounded in source
1
Vendor batch invoices received
trigger
“The vendors they use more frequently often send batches with anywhere from 1-300 invoices, 3 times a week”
2
AI invoice data extraction
ai_action
“Nanonets receives invoices from Ryan's team, captures critical information in real-time”
3
GL code auto-population
integration
“looking up the vendor name in the invoice in a GL Code - Vendor Master CSV file, and auto-populating this information in the CSV exported by Nanonets”
4
Human verification
human_review
“verify the information, and import into our current system”
5
Foundation-compatible CSV export
output
“exports the validated information to a standard CSV file format that can be easily ingested by the Foundation Accounting software”
6
Foundation records update
integration
“The GL code key, present in the CSV containing the extracted information is matched against the GL code on Foundation to update records”
Reported outcome

Ryan can now click and drag a batch of invoices into Nanonets, verify the extracted information, and import it into Foundation, replacing a process that previously took an entire day per batch.

Reported metrics
Manual data entry time per batch setup to 15-20 hours of an employee's time
Annual manual data entry hours1040 man hours (~4 months a year)
Time to process a single batch manuallyan entire day to process one by one
Reported stack
NanonetsPre-trained Invoice ExtractorFoundation Construction Accounting software
Source
https://nanonets.com/customer-success-story/acm-services
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Ryan can now click and drag a batch of invoices into Nanonets, verify the extracted information, and import it into Foundation, replacing a process that previously took an entire day per batch.

What tools did this team use?

Nanonets, Pre-trained Invoice Extractor, Foundation Construction Accounting software.

What results were reported?

Manual data entry time per batch set: up to 15-20 hours of an employee's time; Annual manual data entry hours: 1040 man hours (~4 months a year); Time to process a single batch manually: an entire day to process one by one (source-reported, not independently verified).

How is this accounts payable AI workflow structured?

Vendor batch invoices received → AI invoice data extraction → GL code auto-population → Human verification → Foundation-compatible CSV export → Foundation records update.

Related accounts payable cases