Document AI in SnapApp

Document AI is a feature in SnapApp that helps you automatically read, understand, and process documents. Instead of manually reviewing files, Document AI analyzes documents for you and extracts useful information.

It converts unstructured data (such as PDFs, images, or handwritten files) into structured data (clearly defined fields and values). This structured data can be stored in databases, making it easy to search, analyze, validate, and use in workflows.

Document AI uses machine learning and Google Cloud technology to create scalable, cloud-based document processing solutions that work reliably even as document volume increases.

Common examples include extracting data from:

• Images
• PDFs
• Scanned documents
• Handwritten text

Document AI is built on Google’s Vertex AI platform and uses generative AI to provide end-to-end document processing. This means documents are processed automatically from upload to extraction without requiring deep machine learning knowledge from the user.

To learn more, see the Document AI overview

Table of Contents

• Document AI Parser
• Types of parser
• Built-in Parsers
• Specialized Parser
• Specialized Parser List
• Custom Parser
• How to create a custom Document AI Parser?
• 1. Create
• 2. Define
• 3. Upload
• 4. Train
• 5. Test
• 6. Deploy
• How to import a Document AI parser?
• How to do real time extraction and validation using Doc AI?
• Available Settings
• Pre-Conditions
• 1. Create a custom field for real-time extraction
• 2. Data extraction and validation
• How to validate rules using Document AI?

Document AI Parser

A Document AI parser acts as a bridge between a document file and a machine learning model. Its main responsibility is to read documents and convert their content into meaningful data.

The parser understands the document structure and performs tasks such as:

• Classifying documents (identifying what type of document it is)
• Splitting documents (separating multiple documents within a single file)
• Extracting data (pulling specific fields like names, dates, or amounts)
• Analyzing content (understanding layout, sections, and entities)

Parsers are essential because they turn raw documents into structured and usable data inside SnapApp.

Types of parser

SnapApp supports different types of Document AI parsers based on the document format and business use case.

Built-in Parsers

• Document OCR Uses Optical Character Recognition (OCR) to read text from documents and images. It also captures layout information and automatically improves image quality by correcting skewed or unclear scans. https://cloud.google.com/document-ai/docs/enterprise-document-ocr

• Form Parser Designed for structured forms. It identifies and extracts key-value pairs, table data, and checkbox values from documents such as applications and surveys. https://cloud.google.com/document-ai/docs/form-parser

• Layout Parser Detects document structure such as paragraphs, headings, and content blocks. It helps understand how information is organized within the document. https://cloud.google.com/document-ai/docs/layout-parse-chunk

Specialized Parser

Pre-built parsers optimized for specific document types and industries, such as financial, identity, or tax documents.

Specialized Parser List

Custom Parser

Custom parsers are used when built-in parsers cannot extract the required information. These parsers are trained specifically for your document type and business needs.

• Extractor Extracts raw text, layout information, and normalized entities from documents. https://cloud.google.com/document-ai/docs/extracting-overview

• Classifier Automatically categorizes documents into predefined groups. https://cloud.google.com/document-ai/docs/custom-classifier

• Splitter Identifies document boundaries and separates multiple documents within a single file. https://cloud.google.com/document-ai/docs/custom-splitter

• Summarizer Generates concise summaries for both short and long documents. https://cloud.google.com/document-ai/docs/custom-summarizer

To learn more, see the Parser List

How to create a custom Document AI Parser?

Follow these steps carefully to create and deploy your own custom parser:

1. Create

• Open the Google Cloud Console
• Navigate to the Document AI dashboard
• Click Create Processor to begin
• Select Custom Processor to define your own parser

2. Define

• Specify the document formats the parser should support
• Define entity types and extraction rules using JSON or YAML
• Configure any required pre-processing or post-processing logic

3. Upload

• Upload a sample document that represents the documents you want to process
• This document is used for both training and validation

4. Train

• Click Train to start the learning process
• Wait until the training process completes successfully

5. Test

• Upload a new document
• Verify that the extracted fields and values are accurate

6. Deploy

• Click Deploy to make the parser available
• The parser can now be used within SnapApp workflows

Note: Make sure you have the required Google Cloud permissions.

To learn more, see Document AI Create Processor

How to import a Document AI parser?

This feature is mainly used for Custom Document AI Parsers that have already been created in Google Cloud.

1. Go to Settings from the top bar
2. Select Import Doc AI from the Data menu
3. Enter the Parser Name to identify the parser
4. Enter the Expected Entities that should be extracted
5. Click on Import.

Example:

How to do real time extraction and validation using Doc AI?

Available Settings

Pre-Conditions

• The object must have Elevated Admin Access
• Track Attachments must be enabled to allow document uploads
• A Document AI field must be configured with validation options

Example:

1. Create a custom field for real-time extraction

Use the formula below to dynamically extract data from the uploaded document:

=LOOKUP(CONCAT("gs://",[[docAI_type_fieldname]]), "object_of_selected_parser", "_gcs_uri", "data_extraction_field_name")

Output :

Extracted value (from data_extraction_field_name)

2. Data extraction and validation

1. Go to Settings → Objects
2. Open the required object
3. Click Preview to test the setup
4. Add a new record
5. Upload the document and save the record
6. The system automatically extracts and validates the data

Examples:

How to validate rules using Document AI?

Validation rules are used to compare extracted data against predefined conditions to ensure accuracy.

1. Enable Document AI validation and publish the view
2. The Validity column becomes visible
3. Go to Edit Fields
4. Select the field that needs validation
5. Add a validation expression that defines the condition
6. Define an error message to display if validation fails
7. Save the configuration

Examples:

If the condition fails:

Your Document AI parser is now ready to extract and validate data in SnapApp.

Thank you for following these steps to configure your SnapApp components effectively.

For help or support, email us at snapapp@bluevector.ai.