Chatbot

Overview

By setting common phrases that customers might use to ask about specific knowledge base articles, such as “How much does this product cost?” or “Is this product free?”, you can help the bot respond more accurately to customer inquiries. A variety of example phrases helps the bot better understand possible customer questions and thus provide more accurate answers.

Please note that although the system will attempt to provide the most appropriate answers based on the configured user phrases and knowledge base information, these search results are for reference only and are not guaranteed to be completely accurate. Therefore,in actual conversation scripts, we recommend still offering customers some alternative communication methods, such as an option to reach a live agent, to ensure that in special situations or problems, service can be provided to customers more effectively.

Improve hit rate and accuracy

The knowledge base search feature uses natural language processing (NLP) technology to attempt to understand and process keywords in customer messages. It aims to precisely match customer queries with relevant content in the knowledge base to quickly provide answers.

However, because different businesses or industries may have specific terminology or usage habits, the built-in lexicon may not fully cover all proprietary terms. Therefore, in order tofurtherimprove the hit rate of knowledge base searches, we encourage you to add or update professional terms and keywords according to your business needs.

By personalizing your vocabulary library, you can make the knowledge base search function more accurate, thereby improving the efficiency and satisfaction of customer service.

Lexicon

You can add more proprietary terms by clicking “Lexicon” under the “Chatbot” category in “Apps,” which helps prevent user messages from being misunderstood. For example, if a customer types “how to short sell,” because “short sell” may be a specialized term, you can add a “Stocks” category and then add the term “short sell” in that category to improve understanding of such specialized terms. This setting helps ensure messages are easier to understand.

User phrases

When a customer asks a question, the system will match based on previously configured user phrases, knowledge base article titles, and content. For example, if we configure the following user phrases:

1. When will it be received

2. How many days until the item arrives

3. How many days until the parcel arrives

4. Will the parcel arrive within three days

5. Is there a scheduled delivery date and time for the parcel

6. When will the parcel be delivered

When a customer asks related questions such as “When will my package be delivered?” or “I want to check when my package will arrive,” the system will compare the customer’s question with keywords from these user phrases like “package” and “delivery.” The system uses specific algorithms and fuzzy search to evaluate match scores and will respond with the knowledge base article with the highest match score.

That is, even if the customer’s question does not exactly match the user phrases we configured, as long as there are related keywords or similar expressions, the system can still find and provide relevant answers.