AI Sentiment Branch
Emotion-Aware Routing

What is the AI Sentiment Branch module?

The AI Sentiment Branch module detects the emotional tone of a visitor's message and routes the workflow to an appropriate branch based on that sentiment. It can either ask the visitor a fresh question or silently analyse the most recent user message already in the conversation - making it highly flexible for different workflow positions.

Sentiment routing enables workflows that respond empathetically. A visitor expressing frustration can automatically be prioritised for human escalation, while a satisfied visitor continues through self-service. This reduces escalation rates by routing the right people to the right paths.

How it works

1. The module either asks the visitor a new question (active mode) or analyses the last user message in the conversation (silent mode via analysis_source: last_user_message)
2. The visitor message is sent server-side to the AI via hybridAI.php
3. The AI classifies the sentiment and returns the appropriate next_step_id
4. An optional acknowledgement message is displayed before the workflow continues
5. The workflow branches to the matched sentiment step

Configuration options

• analysis_source - Set to last_user_message for silent mode, or leave unset to prompt the visitor with a new question
• display_message - Optional question or acknowledgement text shown to the visitor
• sentiment branches - Configure route destinations for each sentiment category (e.g. positive, negative, neutral, frustrated)
• fallback_step - Route used if sentiment cannot be determined

Silent vs active mode

Silent mode (analysis_source: last_user_message) analyses the most recent user message without showing any prompt. This is ideal for placing the module after a form field or button response to react to how the visitor expressed themselves, not just what they selected.

Active mode asks the visitor a direct question - such as "How are you feeling about your experience so far?" - and analyses the response. This is better when you want an explicit sentiment check at a workflow transition point.

When to use AI Sentiment Branch

• After an offline message step - detect whether the visitor is frustrated to prioritise the callback
• Mid-workflow after a support query to decide whether to escalate or continue with AI
• After a form submission to gauge customer satisfaction in real time
• As part of a retention workflow to identify unhappy customers before they leave