Analytics

View analytics of your AI activities on Analytics. With Analytics, you can see how your AI performs and how your customers engage with your subjects.

Metrics

At the moment, Mind Expression provides analytics for API Calls, and Popular Subjects counts.

AI Activities Report

You can view a record of chat logs as well as basic chat summary statistics in the AI Activities Report panel.

How to extract the reports:

  1. Select the Scope that you would like to extract analytics from.

  2. Select the date range of the desired data.

  3. Click Export Report. A zip folder containing two files will be downloaded: chat_history.csv and chat_history_statistics.csv

chat_history.csv

This file contains all chat logs from both human users and the AI responses. It includes the following fields:

  1. Scope Name - scope from which the data is extracted

  2. Scope ID - unique ID assigned to the corresponding scope

  3. Channel - integration channel in which the chat was triggered

  4. Timestamp - exact timestamp in which the chat was entered in the format:

    -yyyy MMM dd HH:mm:ss.SSS zzz

  5. Conversation ID - a unique ID assigned to one whole conversation, which begins once the CAI is triggered and ends once the user terminates the session or experiences a session timeout.

  6. Query ID - a unique ID assigned to each query within a conversation

  7. Chat - the exact output or log of the chats from the user and the AI

  8. Is Query - TRUE if the chat is from the human user, FALSE if the chat is a response from the AI

  9. Subject Recognition - the specific subject name that was triggered by the chat

  10. Is Fallback - TRUE if the chat is a fallback, FALSE if the chat is not a fallback

chat_history_statistics.csv

This file contains basic summary statistics of the chat logs that were extracted from the chat history. It contains the following metrics:

  1. Date Range - time period covered by the data

  2. Total number of transactions - total number of inquiries that have been responded/answered with substantive information to the users of the Customer

  3. Total number of queries - Number of chats in the history log that are entered by the user

  4. Number of unique chats - total number of unique conversations that have been completed, i.e, a chat that was triggered and terminated.

  5. Average number of transaction per chat - average number of chats that a user enters within one whole conversation

  6. Chat Duration - length of time a user takes to complete a whole conversation from start to finish. The percentile breakdowns are also presented.

  7. Number of fallbacks - number of chats that are not recognized by the conversational AI and thus resulted to fallbacks

  8. Subject Activation Counts - a breakdown of the number of times each subject has been activated

API Calls

You can see the total API calls during the selected timeline. Use the drop downs to choose the time you want to know the analytics and the time format for the data display.

The Popular Subjects section authorizes you to browse the overall tally of your scope. It allows you also to see how many of your users have been able to activate your Subjects.

Last updated