Sample Definitions for Brand Insights

CSV 1:Conversation Level (agent_feedback_conversation_combined)

business_id

Unique identifier assigned to each bot. (IVA)

4057 | 3010 | 3614

Business_Name

Unique bot name.

upstoxwa | upstoxweb | havells

claim_name

Used as a filter to distinguish whether the bot or an agent performed an activity. In the case of an agent, the agent's name will be displayed.

bot | anurag_b | soumyap

coll_id

Unique identifier for all conversations by a user with a business. Example: If User A has 5 conversations with a bot, all of these conversations will have the same Collection_ID.

122852559

conversation_identifier

Unique identifier for each conversation between a user and a bot. It is a combination of Collection_ID and the number of conversations between the user and the bot separated by an underscore sign. 

Example: If User A has 5 conversations with a bot, the Conversation_Identifier would be: 122852559_1, 122852559_2, ... upto 122852559_5

device_platform

This is the channel or platform where this conversation took place.

Android SDK | WebSDK | Whatsapp

disposition_reason

Gives the reason why an agent was required. Only applicable to chats that are transferred to an agent. This is the text entered by the agent while closing a chat.

Customer Profile / KYC | Spam chat | Sales Enquiry

end_to_end_gogo_chat

Whether the bot handled the entire conversation from start to end. True (1) means agent handover was not required and chat was 100% automated.

1 or 0

feedback_comment

This is the comment left by the user on the completion of the chat/conversation as feedback.

I was able to clear all of my doubts | Was not able to help | Thanks

feedback_metric

Used as a filter for the feedback rating to distinguish between the two types of feedback metrics: CSAT and NPS.

CSAT | NPS

language_script

The script in which the messages were shared by the user. Example: the 'Mujhe help chahiye' script would be English even though the meaning is comprehended in Hindi.

English | Hindi

source_language

The source language of all messages in a conversation by the user. Example: 'Mujhe help chahiye' source language will be Hindi.

English | Hindi | Marathi | Telugu

created_at

Timestamp when the chat got closed. Includes both date and time at minute precision.

-

agent_resolution_time_wqt

The total time it took to solve the user's query from start to finish but does not include the queue time i.e. the time it took to transfer the chat from bot to an agent. Will be applicable even if the chat was handled end to end by a bot. Claim_name can be used as a filter here to distinguish between bots and agents.

150

avg_response_time_wqt

The time taken by the bot or agent to respond to the user's message. This does not include the time taken to transfer the chat to the agent. Claim_name can be used as a filter here to distinguish response times between bots and agents.

20

conversation_resolution_time

Total time taken for the conversation to end in seconds.

175

feedback_rating

Numeric rating given by the user. Out of 5 if CSAT has been activated and out of 10 if NPS has been activated for the client

4 | 9

first_agent_response_time_wqt

The time taken in seconds for the first response message by the bot or by an agent. Claim_name is used as a filter to distinguish between the first response time between bot and agents.

5

conversation_no

Shows the nth number of conversations between the user and the bot. 

Example: For any Conversation_Identifier 122852559_3, the value after the underscore is the number of conversation by the user with this bot

10

total_assistant_messages

Total number of messages sent by agents.

5

total_gogo_messages

Total number of messages sent by the bot.

7

total_flows

A total number of queries that were identified in a conversation. 

Example: One conversation can have multiple queries like KYC, Product discovery, purchase, etc. 

2

total_queue_time

Time taken in seconds to transfer the chat to an agent.

10

total_resolved_flows

The number of queries that were resolved out of the total flows.

1

total_user_messages

Total number of messages sent by the user.

12

total_wait_time

Time in seconds where the user was idle during the conversation.

100

conversation_identifier

This field can be used to join both CSVs if needed. It acts as the foreign key at the message level and the primary key at the conversation level.

122852559_3 | 133468830_2

created_at_with_seconds

The time frame in seconds precision when this message was created and recorded.

-

is_freeform

Whether a message is freeform or not. 

Example: 'I need help. Please transfer to an agent' would be a freeform message. But if any message has been replied to from the quick reply option, it will not be a freeform message.

1 or 0

Skill

It is the Skill or the Step that this message was a part of.

-

Skillset

The skillset this message was a part of.

-

business_via_name

upstoxwa | upstoxweb | havells

is_break_message

Whether a user message caused Bot Break.

1 or 0

is_handover_message

Whether a user message caused Agent Handover.

1 or 0

coll_id

Unique identifier for all conversations by a user with a business. 

Example: If user A has 5 conversations with a bot, all of these conversations will have the same Collection_ID.

122852559

message_id

Unique identifier for each message in a conversation

-

business_id

Unique identifier assigned to each bot (IVA).

4057 | 3010 | 3614