Overview

The Word Mapper is a specialized tool within the Workflow module designed to enhance the bot's natural language understanding (NLP) by creating a custom dictionary of synonyms and equivalences. Its primary function is to allow an administrator to map various linguistic terms, acronyms, or jargon that users might input to a single, standardized term that the bot can easily recognize.

This feature is crucial for improving the accuracy of intent recognition. For example, if your internal systems and bot flows are built around the term "INTT," but your customers frequently ask about "interest rate," Word Mapper allows you to create a mapping so the bot treats both phrases as identical. This ensures that regardless of the user's phrasing, they are directed to the correct conversational flow.

Essentially, Word Mapper acts as a translation layer, bridging the gap between how real users talk and how the chatbot's knowledge is structured.

How to Use Word Mapper

This guide explains how to access Word Mapper and add new term mappings.

2.1 Navigating to the Word Mapper Page

  1. From the main console menu, navigate to Workflow -> Word Mapper.

2.2 Adding a New Mapping

  1. On the Word Mapper page, you will likely find an interface for managing your list of mapped terms. Click the button to add a new mapping (e.g., "+ Add Mapping" or similar).

  2. You will be presented with fields to define the relationship:

    • Term/Keyword: Enter the word or phrase the user is likely to type (e.g., "interest rate").

    • Mapped Term: Enter the standardized term that you want the bot to recognize (e.g., "INTT").

  3. Click Save or Add to add the new mapping to the bot's repository. The bot will now treat any instance of the "Term" as if the user had typed the "Mapped Term."

2.3 Managing Existing Mappings

  1. The Word Mapper page will display a list of all existing mappings.

  2. From this list, you should be able to Edit or Delete any mapping that is no longer needed or needs to be corrected.

Common Examples and Use Cases

  • Handling Acronyms and Jargon:

    • Map the internal term "INTT" to the user-facing phrase "interest rate."

    • Map "KYC" to "Know Your Customer verification."

  • Accommodating Slang and Informal Language:

    • Map "acc balance" and "bal" to the official intent "check account balance."

  • Correcting Common Misspellings:

    • Map "lon" or "laon" to "loan" to ensure users who make typos are still understood.

  • Standardizing Product or Service Names:

    • If you have a product named "SuperSaver Account," you could map terms like "savings account" or "saver account" to it to guide users correctly.

Best Practices

  • Be Specific: Create mappings that are unambiguous. Mapping a very general term like "help" could interfere with multiple intents, so be precise.

  • Test Your Mappings: After adding a new mapping, go to the bot preview and test it. Type the term you added (e.g., "interest rate") and verify that it triggers the correct flow associated with the mapped term (e.g., "INTT").

Common Mistakes

  • Creating Overly Broad Mappings: Mapping a common word like "account" could lead to confusion. If a user types "I have an issue with my account," the bot won't know if they mean a savings account, checking account, or user profile.

  • Mapping to a Non-Existent Term: Mapping "interest rate" to "INTT" is useless if you haven't built a flow or intent that is triggered by the word "INTT." Ensure the destination term is a valid trigger for a conversation.

  • Assuming Two-Way Mapping: Creating a mapping from "Term A" to "Term B" does not automatically create a mapping from "Term B" to "Term A." The relationship is one-way.