👓 AI Services

Overview

AI services in AIForged are specialized processing engines—either first-party or third-party—designed to perform targeted automation and data processing tasks. These services can be flexibly connected using built-in utilities and verification engines to form robust document and data workflows.

Types of AI Services

AIForged offers a variety of AI services, each tailored for specific functions:

• Scrapers Extract documents or data from specified sources (e.g., mailboxes, folders, cloud storage).
• Classifiers Identify and classify the type or category of each page or document.
• Extractors and Verifiers Extract targeted information from documents and verify its accuracy and validity.
• OCR (Optical Character Recognition) Convert images or scanned documents into machine-readable text.
• Object Detection Identify and locate objects within images.
• Speech Recognition Convert spoken words in audio files into text.

Tip: You can combine different AI services to automate complex, multi-step processing scenarios.

AI Services Structure

AI services in AIForged can be configured in several ways to best fit your workflow needs:

• Sequence Services execute one after another in a defined order.
• Independent Services operate completely independently from one another.
• Parallel Multiple services run simultaneously, each working on the same or different data.
• Tandem Services collaborate closely, with outputs from one service feeding directly into another.

Understanding these structural options is key to leveraging the full power and flexibility of AIForged.

AI Service Operations

AI services can be connected in any order, allowing for customized processing pipelines.

Individual AI Services

• Operate independently.
• Do not require input from any other service.
• Suitable for simple, single-step automation tasks.

Dependent AI Services

• Rely on additional functionality (such as OCR or custom logic).
• Can be configured to run:
  • Before Processing: Pre-processes input to enhance data quality.
  • After Processing: Post-processes results for validation or further enrichment.

Example:\ If the Microsoft Document Intelligence service does not provide the desired accuracy, you can integrate custom code before or after its execution to improve results.

Tip: Dependent AI services are set up using the same process as standard AI services.

Verification AI Services

Verification AI services are triggered automatically based on specific conditions—such as when a field’s confidence score falls below a set threshold.

• Can be configured and managed within the Rule Engine.
• Enable automated double-checking and human-in-the-loop escalation for improved data quality.

Tip: Use Verification AI services for critical data points where accuracy is essential.