ADP Product Manual (SaaS)
输入“/”快速插入内容
ADP Product Manual (SaaS)
ADP Product Manual (SaaS)
用户2938用户2938
用户3485用户3485
用户5952用户59523月2日修改
1.
Agentic Document Extraction
1.1
Key Concepts
1.1.1
App
An instance created in ADP to implement specific document extraction tasks. You need 3 steps to build an app: Create → Configure → Test. Apps can then be integrated into external systems via API.
1.1.2
Schema
The rules that define how extraction is performed, including:
◦
Fields: field name, type, and description.
◦
Overall extraction rules: natural language descriptions of overall rules.
1.1.3
Field
Specific data items to be extracted from documents.
1.2
Quick Start
It only takes 3 steps to quickly set up a document extraction app.
1.2.1
Step 1: Create an App
Log into ADP, and click Extraction in the left navigation menu. Upload the document you want to extract in the right-hand panel to create an app.
You can upload up to 10 documents per batch. Each document can be up to 100 pages, 50MB, and in the following formats:
Document Type | Format |
Image | .JPEG, .JPG, .PNG, .BMP, .TIFF |
PDF | .PDF |
Microsoft Excel | .XLSX, .XLS |
Microsoft Word | .DOCX, .DOC |
Alternatively, you can test with sample documents provided by the platform.
1.2.2
Step 2: Build a Schema
Three methods are available:
a.
Generate a schema from a prompt
b.
Generate a schema from smart suggestions
c.
Build a schema from scratch
1.2.3
Step 3: Test and Call API
While configuring the schema, you can click Test at any time to verify the extraction results on the displaying document. Test results can be copied or exported. Detailed instructions can be found in the Testing Schemas section. Once completed, click Save to update the latest schema version.
1.3
Out-of-the-Box Agents
Upload Invoice/receipt/purchase order documents via quick access entry, enabling rapid extraction without manual configuration.
No configuration needed, Click and Use