📄️ Stats & Conversations
CrawlChat provides useful stats and analysis so that you get to know how the chatbot is performing across all the channels. Most of the analysis on CrawlChat is based on the score given to the context (RAG chunks) it fetches for the questions. This score varies from 0 to 1 where 0 means the fetched context is irrelevant and 1 represents high relevance for the given query. Here are few stats it shows from the dashboard. All the information is for last 7 days
📄️ Answer correction
It is understandable that the AI can go wrong at times. Mostly, we can idently them using the user 👎 message ratings or low rating messages. You can go through such messages and if found that the answer provided is inaccurate or incomplete, you can hi Correct it button which opens up a text box where you can enter the expected answer or any information in general.
📄️ Data gaps
One of the capabilities of CrawlChat, is to identify the Data Gaps in your documentation for the queries being asked by the users across the channels. A Data Gap can be missing information or partial information.