The new integrated Intent Discovery module is designed to help you auto-extract popular intents from previous user conversations. It helps automate the creation of intents and provides insights for the conversation design. Upload your existing information sources like previous chat transcripts, support tickets, and knowledge base to discover the popular intents. It reduces the time and effort to build a virtual assistant and leads to the success of your Conversational AI Journey. This is a beta feature and is available only for the English language and for Enterprise users.
Intent Discovery uses the existing information sources like your previous chat transcripts, support tickets, and knowledge base to train your bot. The chat transcripts and support tickets act as base information to select the top use cases. It also can be used as training data so that you don’t have to individually add intents to the bot. It eliminates the manual effort and is time saving.
You have to upload the transcripts and create a project specific to the use case. The following sections explain how to create a project, upload chat transcripts to extract the conversations, train the utterances, and how to get the insights of the project.
Create a Project
- Go to Build > Tools > Intent Discovery.
- Click the button Create Project to create a new project.
- Enter the Project Name and the Project Description and click Proceed to successfully create a project.
Conversational insights like Total files uploaded, Conversations Found, Intents Found, Entities Identified, Total Intents added to the bot, and Entities added to the bot are displayed for each project.
The following different statuses of the project are displayed under the Status column:
- Conversation In Progress
- Conversation Extracted
- No Files Uploaded
- Intent Extraction In progress
- Intent Extracted
- <#number of> Intents Added
- <#number of> Intents Trained
- <#number of> Intents Added,
Intents Trained
Note: You can edit the description or the project name. The platform also allows you to delete the project if required.
Upload Transcripts
Once a project is created, you can upload one or more chat transcripts. The platform provides an option to download a sample transcript file (CSV format).
Note: The sample transcript file has the standard CSV template, in which the transcripts must be uploaded to the platform. The file will not be extracted if it deviates from the format.
- Click on any created project. You will go to the Transcripts tab.
- Click the button Upload Transcript to upload the chat transcript file.
- Upon successful upload, all the transcript files are displayed under the Transcripts tab. The upload status is shown as below.
Note: You can upload one or more CSV files.
The following details of the uploaded transcript files are captured and displayed:
- File Name – Name of the transcript file
- Upload Time– Time taken to upload the file
- Number of Conversations – Unique conversations between the agent/bot with end users.
- Average Messages – Average number of messages exchanged between agent and end-users per conversation
- Status – Status of the uploaded file.
- Conversation Extraction In progress
- Conversation Extracted
- Invalid File
- Upload Failed
- Intent Extraction In progress
- Intent Extracted
- Intent Extraction Failed
The following statuses are maintained for each of the files that are being uploaded:
- The platform also displays the following insights:
- Number of Conversations – Unique conversations between the agent/bot with end users.
- Average Message – Average number of messages exchanged between agent and end-users per conversation.
- In the Transcripts page, click the button Extract Intents and click Confirm on the pop-up to proceed with extraction.
- The intents get extracted and the status can be seen across the file names as shown in the following screenshot.
- Once the extraction is completed, click on the button Review Intents.
- You can see the list of extracted intents under the Intents tab.
- Details of the mapped intents can be viewed on clicking the View More link in the above screenshot.
- Click on any extracted intent to view all the influenced utterances that have led to intent identification.
- Click the View icon to see the transcripts of the utterances.
- Create a new Dialog Task or FAQ
- Train them for an Existing Dialog Task or FAQ
- Change the status of the utterances
- Under the Influenced Utterances, select the utterances for which you want to add a new Dialog Task.
- Select Dialog Task option from the Add Intent drop-down.
- A new dialog task creation window is displayed.
- You can add or modify the Intent Name that has been extracted from the conversations and add all other details in the New Dialog Task panel.
- Select the utterances to be added as alternate questions and click FAQ from the Add Intent drop-down.
- Add FAQ panel is displayed, where you can enter all the relevant details
- To map an existing Dialog Task or FAQ, click the Train drop-down.
- Click the Dialog Task option to choose the matching intent for the selected utterance(s).
- Click the button Add Utterances to add the intent for the utterances.
- Similarly click the FAQ option from the Train drop-down to add the matching FAQ for the selected utterance(s).
Delete Transcripts
The uploaded transcript file can be deleted if the file is not required anymore, in cases where the upload was incorrect or if the conversation has been extracted and the file is not useful.
View Conversations in the Transcript
Upon successful upload of the transcript, you can view all the conversations and the chat messages exchanged between the end user and the VA/agent.
The platform displays the conversation session summary, which includes the count of the VA messages and user messages.
Note: If a conversation is really not useful and is not contributing to the intent and entity extraction, you can delete the specific conversation. Deleting a conversation also impacts the conversation count and average message count in the columns.
Extract Conversations
The platform can extract the conversations from the uploaded files. After validating all the conversations within transcripts, the next step is to extract intents, entities and training data from the conversations without spending time in segregating them manually.
Note: You should be able to extract the intents only once during transcription, for a project. If you need to extract intents from the same file again, then you need to create a new project. Intent extraction may take between 3-5 hours depending on the number and nature of the conversations.
Follow these steps to extract the intents from the transcript file:
All the extracted intents are listed with conversation count, count of entities, and the count of utterances that are extracted from the uploaded files. You can add tags to the extracted intents. You can add more than one 1 tag and also remove these tags.
The platform displays the list of utterances, mapped intents and status of the utterances (Reviewed/Yet to Review). The platform also displays details of all the intents and FAQs that are added or trained for the extracted intents.
Train the Utterances
Once the utterances are listed you can select one or more utterances and perform the following actions:
Create New Dialog Task
In a scenario where the utterances cannot be mapped to any of the existing intents, you can create a new intent. Follow these steps to add a new intent (DialogTask or FAQ) for the utterance:
Note: You have the option to delete the utterances if they are not valid for the intent.
The More Options feature will have the same options as in dialog creation.
The platform also displays the entities it has identified from the utterances that were selected to create dialog.
Note: You can view the utterances that influence the entity formation and also have the option to delete the entities.
Add New FAQ
You can also train your utterances to create new FAQs. Follow these steps:
Note: You need to add the primary question and the bot responses to create the FAQ.
Once a dialog task or FAQ is added, the platform displays the Mapped Intents for each utterance.
Note: The platform displays all the FAQs and Dialog tasks an utterance is mapped to.
Train Utterances for the Existing Intents
The following steps explain how to add an existing intent to an utterance or multiple utterances using the Train drop-down.
Change Status
You can modify statuses of the utterances that have been extracted from the files. Once the utterances are added or trained to an intent, a message is displayed to change the status. The status of all the utterances is Yet to review by default. You can change the status of an individual utterance from Yet to review to Reviewed and vice versa.
Note: You can also use the Status drop-down provided in the top right corner of the page to change the status of one or more selected utterances.
Upon completion of training, you can see the details of the trained utterances in the Mapped Intents pop-up. You can also see the tags added based on the status of the utterances, under Tags column.