Docs
  • Overview
    • 👋Welcome to Orimon AI
  • Product Guides
    • 🤖Create A Bot
      • đŸĨLeads Generation & Support Bot
      • đŸ’ŗE-Commerce Specialist Bot
    • đŸ“ĒFlag and Train
    • âš™ī¸Setting up Bot Appearance
      • 🎨Setting Theme Color
      • đŸ’ŦSetting Message Bubble
      • 🎓Setting Bot Header
      • đŸ—¯ī¸Setting Trigger Icon
      • 📩Setting Welcome Message
      • â—€ī¸Setting Fallback Message
      • đŸŗī¸Disable Branding
      • đŸ—’ī¸Setting Lead Form
      • âš ī¸Setting Up Conversation Triggers
      • ⚡Adding Quick Replies
      • 📎Custom Query Placeholder
    • ✨Setting Up Conversational Leads
    • 🙋Setting up Live Chat
      • âš™ī¸Configuring Live chat
      • đŸ—Ŗī¸Transferring To Live Agent
      • đŸ’ģLive Chat Agent Screen
      • đŸ›Ŗī¸Live Chat Process
      • 🕔Operational Hours
      • âšī¸Disable AI
      • 👉Join Active Chat Room
    • đŸ’ŊBot Model
    • đŸŒĄī¸Bot Temperature
    • 🛑Quota Limits
    • 🎭AI custom prompt
    • 🔑Bring your openAI key (BYOK)
  • Fundamentals
    • 🧑Inviting Members
    • âŦ‡ī¸Install the Dashboard App
    • đŸĻBilling and Subscription
      • 🔔How do I subscribe?
      • 🔋How to buy additional quota for my plan?
      • 📈How do I upgrade my plan?
      • ⛔How do I cancel my plan?
      • â™ģī¸How do I renew my plan?
  • Integrations
    • đŸĒWebhook
    • ⚡Zapier
    • đŸ›ŗī¸Pabbly
    • đŸĒ¸Hubspot
    • đŸ‘ĨSalesForce
    • 📗Google Sheets
  • Channels
    • đŸ—Ŗī¸Setting up Channels for Messages
    • 📩FaceBook Messenger
    • đŸŸĸWhatsApp
      • đŸ’ŗAdd Payment Method
    • đŸĒ›DIY Website Building Tools
    • đŸŸĻTelegram
    • 📰WordPress
    • 🌐Website Integration
    • ⚡Zapier
    • đŸ›ŗī¸Pabbly
    • đŸ‡Ģ🇴Slack
    • 🎮Discord
    • đŸ‡Ģ🇲MS Teams
  • Templates
    • 🤝Book a Meeting
  • White-Labelling
    • đŸŗī¸Getting Started
    • 🤖Chat-Bot White Labelling
    • 🔰Dashboard White Labelling
    • 📧Email White Labelling
  • developer API
    • 🔍Getting Started with APIs
    • âœ‰ī¸Message API
Powered by GitBook
On this page
  • Step 1 - Flag Responses during testing
  • Step 2 - Flag Incorrect Responses from History
  • Step 3 - Navigate to Training History Page
  • Step 5 - Train Corrected Responses
  • Extras : Bot Training Opportunities

Was this helpful?

  1. Product Guides

Flag and Train

Introducing Orimon AI's Flagging and Training Feature - A Simple Solution for Continuous Learning

PreviousE-Commerce Specialist BotNextSetting up Bot Appearance

Last updated 1 year ago

Was this helpful?

Orimon AI platform offers an intuitive and powerful solution for continuous learning. One of its key features is the Flagging Feature that enables users to flag incorrect responses during test chats or from chat history. The flagged responses are then marked and made available for simple correction in the training section. This feature helps to enhance the chat-bot's knowledge by enabling trainers to update the correct responses and train the bot for better performance.

To use this feature, follow these simple steps:

Step 1 - Flag Responses during testing

During test chats invoked either from testing page on Orimon Dashboard or using test bot links, if you or your team comes across any response that doesn't seem right and needs to be corrected, click on the "Edit & Train" button visible on the respective response message.

Following this, a prompt will open-up asking you to make the necessary modifications to the Chat-bot's output.

Once you have made the necessary changes, you can either "Save" this change for training later or click on "Save & Train" to train the changes now.

Step 2 - Flag Incorrect Responses from History

From chat history, select the message that needs to be corrected and click on the "Flag" button.

Following this, a prompt will open-up asking you to make the necessary modifications to the Chat-bot's output.

Once you have made the necessary changes, you can either "Save" this change for training.

Step 3 - Navigate to Training History Page

All the query-response pairs which are flagged in previous steps will be visible in the training history section and collated together for correction.

Go to the "Test & Train" page once again -> the click on "History" at the top of your chat-bot.

A dialog box will then open up as follows.

Step 5 - Train Corrected Responses

Once all the flagged messages have been corrected, train the bot with the updated responses.Once training is successfully completed you will be redirected to testing screen where you can verify your bot with the updated knowledge you have trained with.

Extras : Bot Training Opportunities

Navigate to Chat History -> Bot Training Opportunities

This section serves as a focal point for refining your bot's intelligence. Here, you'll find interaction pairs flagged for review either by user feedback or our AI observability feature.

How It Works:

  • User Message: The user's initial message that triggered the bot's response. Bot Response: The bot's reply that has been flagged for review.

  • Feedback Source: Indicates whether the feedback was provided by a user or identified through AI analysis.

  • Issue Identified: A brief description of the potential issue with the bot's response. Action Taken: Current status of the interaction, such as updated, pending, or dismissed.

  • Train/Ignore: Options for how to proceed with the flagged interaction, suggesting whether to refine the bot's response or not prioritise it. Use these insights to continuously enhance your bot's accuracy and relevance. Each adjustment is a step towards a more intelligent and reliable chat-bot for your users.

đŸ“Ē