Conversational Data

Important Concepts

The data we use for conversational analysis is generated on the Blip Platform, based on message traffic. The Platform is divided into several “systems” that perform different functions: message traffic itself, human support (desk), active notifications, contacts, etc.

The Blip Platform is based on a protocol that defines the basic structures required for a messaging application. It is called LIME (Lightweight Messaging Protocol) and is available at: https://limeprotocol.org/

An important concept in our data model is the tunnel. Data generated in the router is associated with the user ID of that channel (on WhatsApp it is the contact phone number, for example). For data generated in subbots, users are identified through the tunnelid, which is a masking of the contact identity.

Message

Messages are the contents exchanged by an intelligent contact. They can be conversational flow messages (bot and human) or system messages, such as internal redirects between subbots. Any content sent or received by an intelligent contact, of any type, is considered a message. Text, images, audio, video, and carousels are some examples of messages that can be found in the Messages table. Messages are generated in the context of the Router and Subbots.

Notification

A notification is the status information of messages, generating events each time a message changes state (for example: if it was sent, delivered, read, or if a failure occurred).

accepted: The message was received and accepted by the server.

dispatched: The message was dispatched to the recipient by the server.

received: The recipient received the message. This event is generated by the destination node.

consumed: The recipient consumed (read or processed) the message. This event is generated by the destination node.

failed: A problem occurred during message processing. In this case, the "reason" property of the notification must be present.

The failure event may be associated with a response from Meta or Blip.

Trackings/Categories

Trackings are special types of data generated by actions/events in Blip. They must be created within the Builder and are generated in association with message sending. They are “normalizations” of unstructured data (messages) into a more structured format, making conversational analysis easier.

Ticket

A ticket represents human support, until its conclusion or transfer. It is generated when the client is transferred to the Desk, and a ticket is created in the queue, regardless of the agent's action. It can be closed at various times (in the queue, by the agent, by the client, or due to inactivity).

Contact

A contact represents the user of an intelligent contact. This entity contains various information about the person conversing with a bot, such as ID, name, phone, email, etc.

The Blip Platform allows a wide variety of information in a contact by default, with the possibility to include extras (called "contact extras").

Recording information associated with contacts is directly related to Bot development.

The table linking a contact to its records can be found in the contact object of your share.

Agent

An agent represents the human who will have access to Blip Desk (or any other tool) to assist users if they choose to talk to a human.

Data Dictionary

It is possible to consult the data dictionary of the fields, with the column name, datatype, description, and example of the data, in the file below:

Click to access.

For more information, access the discussion about this topic in our community or the videos on our channel. 😃