AP Capture AI
  • 09 Jan 2025
  • 3 Minutes to read
  • Dark
    Light
  • PDF

AP Capture AI

  • Dark
    Light
  • PDF

Article summary

This article is an overview of the AP Capture AI, which uses Optical Character Recognition (OCR) to automatically complete forms on the Documents to Process page of the R365 Accounting software. 


Overview

The AP Capture AI tool uses Optical Character Recognition (OCR) to automate AP invoice processing, saving valuable time and resources.

Once an invoice has been uploaded in Documents to Process, select the document to open the Draft Transaction. If OCR is enabled, the fields on the Draft Transaction screen are automatically completed by the software. OCR reads and populates fields according to the vendor and item details on the invoice and setup in R365. If a match is not available, users are prompted to fill in required fields. If any corrections need to be made, the fields can be manually adjusted. Alternatively, a user can select the Review OCR Mapping button above the Document Previewer. This opens Review OCR Mapping, where the user can instruct the tool by making mapping corrections directly on the invoice. When corrections are made in the Review OCR Mapping sidesheet, the tool remembers these mapping selections to improve automation for future invoices. Learn more about mapped fields, see Documents to Process: Correcting OCR Mapping.

Users can access and review documents that were created in the past 90 days from the Recently Completed tab of the Documents to Process page.

For more information on processing documents in R365, see Documents to Process Overview.


Security

The permissions listed here are associated with OCR. These permissions can be added to custom user roles or single users. Learn more about managing permissions and custom user roles here.

Permissions to edit OCR mappings are found in the permission tree as follows:

  • System Setup

    • Edit OCR Mappings

The Permission Access report can be used to determine which user roles or users already have these permissions assigned. Learn more about User Setup and Security here.


Setup

OCR looks to the information on the invoice as well as the records saved in R365. OCR does not create new vendors, items, vendor items, payment terms, or gl accounts automatically.

Improve OCR processing by ensuring the following records are entered and up to date:


OCR Processing

Updating fields on the draft transaction sheet when OCR is enabled helps improve the AI by teaching it to recognize and remember the correct mapping for future transactions. The AI learning model enhances its accuracy with each correction, resulting in reduced manual effort as more transactions are processed.

The following fields are read from the invoice/credit memos uploaded:

  • Invoice Number

  • Invoice Total

    • Recognizes negative invoice totals as credit memos, and updates the transaction type.

      • The system updates line item amounts based on the sign of the invoice amount and the existing line item amounts as follows:

        • When the invoice amount is negative and all line item amounts are negative, the line item amounts are converted to positive.

        • When the invoice amount is positive and line item amounts are negative, the line item amounts are converted to positive.

        • When the invoice amount is negative and there is a mix of negative and positive line item amounts, each line item amount is updated to its opposite sign.

  • Line Item Invoice Amounts

  • Invoice Date

  • Due Date

The following fields require existing records in R365 to populate:

  • Location

  • Vendor

  • Payment Terms

  • Vendor Item

  • GL Account

OCR does not recognize handwritten notes unless they are part of a mapped field or column. Handwriting in mapped fields will be processed if it is legible, but handwritten notes elsewhere on the document will not be read.


Review OCR Mapping Sidesheet

Field

Description

File Name

File name of the document.

Submit Corrections

Saves mapping corrections and informs the software to improve accuracy for future mapping.

Document window

Displays an interactive version of the document. Click and drag to draw boxes around items in the document and map them to fields.

Fields

Displays a list of fields along with the data that is currently mapped to each field.

Tables

Data that has been extracted from the document using a table. Tables can be used to extract rows of data from the document. Assign column headers to the correct data type and select the checkmarks in each cell to confirm that the data is correctly mapped.

Zoom

Zoom in or out to better view the document.

Show Grids

If selected, grid lines appear around data that was mapped using a table.

Add Table

Add a new table to extract data from the document.


Was this article helpful?