Configuring Conversation Pipeline

To learn more about the Analyzing Conversation, click here.

Conversation Pipeline is used to group collections of customer interactions. You may have a pipeline for calls, emails, and so on. Each type of analysis ( summarization, tagging, etc.) is then configured specifically for a particular pipeline.

__________________________________________________________________________________

How to configure the Conversation Pipeline?

As an Administrator user, follow these instructions:

1. Once logged in, go to the account's Settings tab in the bottom left-hand corner of your screen  
   
2. Select the CI tab
3. Click the + New  Pipeline on the top right side of the page
   
4. Fill in the information within the New Conversation Pipeline.
   
5. Customize each Pipeline's stages to effectively analyze your customer interactions.
   1. Data Input- This stage is the foundation of your pipeline, where you define parameters that select the specific interactions to be analyzed. To learn more on how to configure Data Query, click here.
   2. Questions- This is designed to help organizations uncover valuable Voice of Customer (VoC) insights by allowing them to ask tailored, open-ended questions about specific customer interactions.
      To learn more about Custom Questions, check these other Help Center articles: Understanding Custom Questions and Configuring Custom Questions.
      
   3. Tagging- This stage lets you add custom tags, intents, and resolution prompts to the pipeline.
      To learn more about Custom Tags, Intent, and Resolution, check these other Help Center articles: Understanding Custom Tags, Configuring Custom Tags, Understanding Intent, Configuring Intent, Understanding Resolutions, and Configuring Resolution.
      
   4. Transcriptions- This stage enables you to select specific words or phrases to follow, tailored to each pipeline. It helps in maintaining professional conduct and ensures content relevance.
      
   5. Prompt- Prompting is a crucial process in guiding large language models (LLMs) like Pathlight to generate valuable responses, involving the provision of clear and concise instructions or questions. This stage allows you to set up Prompt Context, Employee type, Custom Summary, and Sentiment. 
      To learn more about Prompt, check these other Help Center articles: Prompting and Context Best Practices, Understanding Summary, Configuring Summary, Understanding Sentiment, and Configuring Sentiment.

__________________________________________________________________________________

We hope this was helpful! Please submit a ticket here if you have any questions or need further assistance.