Re-train Your Chatbot
Last updated
Last updated
Our chatbot is designed to answer user queries using imported documents, web pages, and other relevant information. While it generally provides accurate responses, there are instances where its answers may fall short of expectations.
To address this issue, we've introduced a new feature called Q&A data source. This feature empowers users to directly edit the chatbot's responses during testing and import them into the chatbot for optimization. Essentially, users can refine the chatbot's responses in real-time, ensuring they are more accurate and tailored to the specific context.
There are three ways to help you retrain your chatbot for more tailored responses:
When you want to add more accurate Q&A pairs, .
When debugging your chatbot, consider .
When reviewing chatbot interactions with users, .
Next, please refer to the detailed operational instructions below. 👇
Navigate to the management panel of the chatbot.
Under the Data Source section, locate the Q&A corpus option.
Click on the import button and input your questions and answers.
Upon completion, import them into the chatbot.
While testing the chatbot's response effectiveness in its interface:
If you find yourself unsatisfied with the provided answer, simply click on the edit button beneath the answer.
Re-edit the answer as necessary and import it back into the chatbot.
Access the Dashboard interface.
Under each question and answer pair, you'll find a small edit button.
Click on it to edit both the question and answer.
After making the necessary adjustments, click save to import the question and answer corpus into the chatbot.
It's important to note that this edit button is specifically meant for retraining the chatbot and will not modify the history of that particular conversation.
