Summary
Planhat includes AI sentiment analysis and Conversation classification
Sentiment is assigned on Conversation level, and rolled up to Company, End User and User (internal team member) level using a time-decay formula to accurately reflect the current sentiment
Conversations are assigned a category from a pre-defined list of 10+ categories
Sentiment and categories can be used in Custom Automations, Health Scores and reporting e.g. Dashboard Pages
Only Conversations within the last 90 days are analyzed when enabling the feature
Sentiment and category classification are charged with Planhat's AI credits
Who is this article for?
Anyone who would like to learn about Planhat's ability to analyze and classify Conversation sentiment and category via AI
Series
Planhat AI: Sentiment and Conversation Classification ⬅️ You are here
Article contents
Introduction
Customer Conversations contains a lot of data and context about customers - emails, chats, calls - that often fall through the cracks. Manually extracting insights from Conversations can be time-consuming.
That's why Planhat now automatically classifies and analyzes the sentiment of customer Conversations - including emails, chats, and call transcripts. This enables your teams to better understand sentiment, urgency, and purpose at scale, while unlocking smarter automation across your workflows.
Each incoming Conversation in Planhat can be tagged with a category, such as Support, Billing & Contracts, or General Enquiries, and evaluated for sentiment (positive, neutral, or negative). The sentiment is rolled up to Company, End User and User level, to unlock deeper insights on which customers have declining relationships, for urgency and priority.
This functionality improves visibility into customer health, team performance, and trending issues - and unlocks automation and reporting.
📌 Important to note
Sentiment and category is assigned on Conversation level, e.g. an entire email thread, and not on each individual email.
What is Conversation sentiment classification?
If "Classify Conversations" is enabled in your Planhat instance (tenant), each incoming email, chat and call transcript is automatically analyzed. Each Conversation is labeled as positive, neutral or negative, and if more emails or chat messages are added to the Conversation thread, the sentiment is re-evaluated.
Sentiment is rolled up to different levels (models) to show trends for specific Companies, End Users or User (internal team members). Each Company, End User and User has a "Sentiment score", which is calculated as follows:
Conversations in the last 90-day rolling window are taken into account
Each Conversation is weighted with time decay, meaning that more recent Conversations affect the scores more than older Conversations
The score is on a scale from -100 to +100, where -100 is very negative and +100 is very positive
The "Sentiment score" shows different colors based the score. These thresholds can be edited in the "Data Model" ("Data") Global Tool, "Fields" tab, for each model (Company, End User and User), by going into the "Sentiment score" field and updating the color conditions. The default values are set to:
Scores ranging from -100 to -30 are displayed as red
Scores ranging from -30 - 30 are displayed as yellow
Scores ranging from 30 - 100 are displayed as green
What is Conversation category classification?
When "Classify Conversations" is enabled in your Planhat instance (tenant), each incoming email, chat and call transcript are automatically tagged with one category based on the content of the Conversation. The category is assigned from a predefined list of 10s of categories (e.g. Support, General Enquiries, Billing & Contracts, Renewals) and is shown in the "Category" field on the Conversation.
Why use sentiment and category classification?
You can use this functionality in Automations to drive intelligent actions, in Health Scores to accurately reflect the sentiment of a customer, or tracking through reporting. Use cases include:
Trigger Custom Automations based on sentiment, e.g. notify CSM when negative Conversation appears for portfolio company
Trigger Custom Automations based on category, e.g. notify Finance Team when a Conversation is flagged with Billing & Contracts
Use sentiment on Company level as a factor in Health Score, e.g. when sentiment declines over a period of time, affect Health Score negatively. Read more about setting up Health Scores here
Track team performance in Dashboard Pages by monitoring sentiment trends by Users or Teams
How to activate sentiment and category classification
To activate sentiment and classification of Conversations in Planhat, navigate to "Settings" Global Tool, and in the "Account Setup" section:
Accept the Planhat AI Terms of Service
Enable the "Classify Conversations" toggle switch
This will analyze the incoming Conversations for the last 90 days based on the time of last reply. Conversations with replies older than 90 days will not be analyzed.
📌 Important to note
Analyzing sentiment and classifying Conversations into categories is charged using AI credits. You can read more about AI credits and permissions here.