Feedloop AI
  • GETTING STARTED
    • Discover Feedloop AI
    • Authentication
      • Sign Up
      • Login
    • Platform
      • Workspace
      • Project
      • People
      • Usage
  • WHAT IS AGENT
    • What is Agent
      • ⚪Manage Agent
    • Basic Agent
      • ⚪Instruction
        • Who
        • What
      • ⚪Training
        • Rules
        • Example
      • ⚪Skill
        • Knowledge
        • Analys
      • ⚪Style
    • Procedures
    • Advance Reasoning
      • ⚪Creating New Agent
      • ⚪Advance Agent Activation
      • ⚪Deleting Agent
      • ⚪Advance Agent Managing Existing
  • AGENT SETTINGS
    • Instruction
      • Who
      • What
      • Additional Instruction
    • Training
      • Chat Rules
      • Retrieval
      • Answering
    • Context
      • Primitive Context
      • Array
    • Tools
      • Knowledge
      • API - Call
        • Manage API Call
        • Basic Configuration
        • Header
        • Body Field
          • Input Binding
            • Page 2
          • Output Binding
        • Testing API Call
        • Trouble Shooting
        • Example Implementation
          • API Call Covid
      • Data Connector
    • Card
      • Intro & Summary Text
      • Data Visualization Card
      • Analyze Card
    • Procedure Basic Info
      • Activation Condition
      • Goal
      • Action
    • Procedures Interaction
      • Procedures Step
      • Procedures Context
      • Procedures Tools
  • AGENT UTILITY
    • Resource
      • Document PDF
        • Preparing your PDF
      • Data Connector
        • Create new data connector
        • Configure dataset, table, and column
        • Using Data Playground
        • Dataset in Data Playground
        • Training in Data Playground
        • SQL generation and modification
        • SQL Execution
    • Apps
      • Chat Widget
      • Web App
    • Testing
    • Monitoring
      • Customer Chat
      • Context
      • Document
  • TASK
    • ⚪Image Analyzer
      • Building Analyzer
      • ID Identifier
    • ⚪Document Analyzer
      • Contractual Document Analyzer
    • ⚪Database Analyzer
    • ⚪Additional Solution
      • Text Sentiment Analyzer
      • Customer Complaint Analyzer
  • For Developer
    • API Chat
      • Chat Initialization
      • Chat Message
      • Chat History
      • Error Handling
    • Embed Widget
      • Embedding the Widget
      • Identity Attributes
      • Widget Attribute Details
Powered by GitBook
On this page
  • By Rules Training
  • Adding Training by Rules
  • Best Practice
  1. WHAT IS AGENT
  2. Basic Agent
  3. Training

Rules

PreviousTrainingNextExample

Last updated 8 months ago

By Rules Training

  • By rules training involves defining specific rules or conditions that trigger particular responses from the AI agent.

  • Users specify the conditions, and the AI agent is trained to follow these rules and produce the desired output when the conditions are met.

  • By rules training allows for more structured and conditional responses from the AI agent.

Adding Training by Rules

  • By Rules Training:

    • Select the "By Rules" training type.

    • Click on the "+ Add Training" button to create a new training rule.

    • Define a condition or pattern that triggers a specific response.

    • Specify the desired output or action that the AI should produce when the condition is met.

  • Save the training rule by clicking the "Save" button.

  • To discard a context, click the "Discard" button instead of the "Save" button.

  • By Rules: This training type supports English input for defining specific conditions and responses. If you prefer to instruct the AI agent with rule-based patterns using English, this is the ideal option for you.

Best Practice

Rules

  • Condition: When the user expresses interest in a specific product but hesitates to make a purchase.

  • Instruction: Offer a limited-time discount or a special offer to incentivize the user to make a purchase, emphasizing the value they will gain from the product.

  • Condition: When the user asks for a comparison between two or more products.

  • Instruction: Provide a detailed comparison chart highlighting the key features, benefits, and pricing of each product, assisting the user in making an informed decision.

  • Condition: When the user inquires about the availability of a product that is currently out of stock.

  • Instruction: Apologize for the inconvenience and offer to notify the user when the product is back in stock. Provide alternative product options that meet the user's requirements.

  • Condition: When the user asks for a demonstration of how a product works.

  • Instruction: Offer to schedule a personalized live demo or provide access to a pre-recorded video demonstration showcasing the product's functionality and key features.

  • Condition: When the user asks for customer testimonials or reviews about a specific product.

  • Instruction: Share positive customer testimonials, reviews, or case studies that highlight the product's success stories and satisfied customers, building trust and credibility with the user.

⚪
Training by Example