DocView Capture
The AWS Textract finder reads invoices and receipts with Textract's expense analysis (AnalyzeExpense): vendor, invoice number, date, totals and the line items. The same finder reads key-value pairs of forms, answers questions about the page (queries), reads tables, or reads passports and driving licences.
Try it
Create a Batch Setup from this template (Batch Setups › New Batch Setup, tab System features), and download its sample documents from the template's card (Download samples). Then: Switch on AWS under System › AI connections and enter the region and keys (or use the server's role). Import invoices.pdf and open a document: the fields and the line items are filled by Textract.
At a glance
| Region | Global |
| Industry | System features |
| Document types | Invoice |
| Workflow | Scan → OCR, separation & classification → Extraction & validation → Index / Validate → PDF generation & export |
| Straight-through processing | No — every batch is checked on screen |
| Separation | keywords at the top of a page |
| Output | searchable PDF with XML data |
| Dates read as | dd/MM/yyyy, d/M/yyyy, dd.MM.yyyy |
Use this template
- Go to Batch Setups and choose New Batch Setup.
- Filter by Global or search for "AWS Textract", select it and choose Next.
- Name the Batch Setup, choose the export folder and create it.
- Try it with a few of your own documents on the Test tab, adjust fields and rules, then publish it — see Test a Batch Setup and Versions and publishing.
Document types and fields
Invoice
Recognised by words such as "invoice", "subtotal", "total due".
| Field | Type | Required | Found by | Checks |
|---|---|---|---|---|
| Invoice number | Text | Yes | entered by the operator | |
| Invoice date | Text | entered by the operator | ||
| Supplier | Text | entered by the operator | ||
| Total | Text | entered by the operator | ||
| Line items | Table | entered by the operator |
Adapt it to your documents
- Add or rename fields and change how they are found: Fields and data extraction, Finders and zones.
- Improve document type recognition: Document classes and classification, Samples and classifier training.
- Send the documents and data to your systems — folders and servers, SharePoint, Google Drive and other cloud drives, S3, document management (FileNet, Alfresco, M-Files, DocuWare), SAP, Dynamics 365, Salesforce, Xero, QuickBooks, SQL, REST or a C# script: Output and export, Cloud and business-system destinations.
- Arrange the data entry screen: Screen designer.