Entity Recognition Skill
Identifying Key Information
The Entity Recognition Skill identifies and extracts specific types of information (entities) from a text. This helps your workflow pinpoint crucial data like names, email addresses, phone numbers, or dates within a sentence.
Extracting Key Details from Customer Free-Text Queries
Imagine a customer types a free-form query like, "My name is John Doe, and my email is [email protected]; I need assistance with loan #L12345." You want your workflow to automatically pull out the name, email, and loan number.
The Challenge
Extracting specific pieces of information from unstructured text manually is tedious and error-prone.
The Solution
Use an Entity Recognition Skill to automatically identify and extract predefined entity types (like PERSON, EMAIL, PHONE_NUMBER) or custom patterns (like loan numbers using a Regular Expression).
Setting Up the Entity Recognition Skill
Locate the Skill: Drag and drop the Entity Recognition Skill onto your Workflow Builder canvas. Place it after the skill that provides the text you want to analyze (e.g., a "Prompt Skill" output or a "Data Extraction Skill").
Configure "Text to Analyze": Link this to the input text (e.g., $input.customerQuery).
Define Entities: You can use pre-supported entities or define custom ones:
Pre-supported Entities: The Skill comes with built-in recognition for common entities like PERSON, EMAIL, PHONE_NUMBER, DATE, LOCATION, etc.
Custom Regex: For unique patterns (like a specific loan number format), you can add a custom Regular Expression (Regex). Provide a name, the Regex pattern, and any flags (e.g., i for case-insensitive). You can even test your Regex directly in the UI.
Understanding the Outcome (Output)
The output will provide a list of identified entities, including their value (the extracted text) and their type (e.g., PERSON, EMAIL, or your custom Regex name).