Configure GoldFinch AI OCR Templates

Purpose

The OCR Template enables GoldFinch to automatically extract structured data from business documents and create GoldFinch records, such as Purchase Invoices.

Using AI-powered Optical Character Recognition (OCR) services, such as Google Gemini, the OCR Template reduces manual data entry, improves accuracy, and accelerates document processing.

Overview

The OCR Template defines how AI extracts data from a document and maps that information into GoldFinch objects.

A typical Purchase Invoice process is:

1. A supplier invoice is received by email or uploaded manually.
2. GoldFinch creates an OCR Request record.
3. The document is sent to the configured AI provider for processing.
4. AI extracts structured data from the document.
5. The OCR Template maps the extracted data to GoldFinch header and line objects.
6. A draft Purchase Invoice is created for review.
7. Users verify or edit the data before posting the Purchase Invoice.

Because the mapping is configurable, the same OCR framework can be reused for many different document types.

How to Set Up an OCR Template

The OCR Template determines how documents are processed, how AI extracts information, and where the extracted data is stored.

need updated screenshot

AI Provider

Select the AI service used to perform OCR.

Examples include:

• Google Gemini
• Other supported AI providers

Model Name (Optional)

Specify the AI model to use for document processing.

If left blank, the default model configured for the provider will be used.

Response Schema

The Response Schema defines the structured data that the AI should extract from the document.

For a Purchase Invoice, common fields include:

Header Fields

• Vendor Name
• Invoice Number
• Invoice Date
• Due Date
• Currency
• Payment Terms
• Total Amount
• Tax Amount

Line Fields

• Item Description
• Quantity
• Unit of Measure
• Unit Price
• Discount
• Tax
• Line Amount

The Response Schema is fully configurable and can be customized for different document types.

Parser Class

The Parser Class converts the AI response into data that GoldFinch can process.

Typical responsibilities include:

• Parsing the AI response
• Validating extracted values
• Formatting dates and numbers
• Handling missing or inconsistent values
• Transforming data into GoldFinch objects
• Preparing header and line records for creation

Different parser classes can be developed for different document types or AI providers.

Target Object

Specifies the GoldFinch object that stores the document header.

Examples include:

• Purchase Invoice
• Expense Report
• Sales Quote
• Custom Object

Line Object

Specifies the GoldFinch object used to store detail records associated with the header object.

For example:

• Purchase Invoice Line

If the document does not contain line items, this field may be left blank.

Extending OCR to Other Business Documents

The OCR Template is designed as a reusable document automation framework and is not limited to supplier invoices.

Examples include:

• Purchase Invoices
• Purchase Orders
• Sales Quotes
• Expense Receipts
• Packing Slips
• Bills of Lading
• Shipping Documents
• Inventory Documents
• Manufacturing Documents
• Contracts
• Custom Business Forms

Because the Response Schema, Parser Class, and object mappings are all configurable, the same OCR framework can support virtually any structured business document without requiring changes to the core OCR engine.