Branch on Customer Sentiment with AI Intent Detection

AI Intent Detection reads a piece of text and sorts it into one of three outcomes: Positive, Negative, or None. The workflow then splits into three matching branches so you can respond appropriately to happy, unhappy, or unclear customers. It is a premium action, so each run has a usage charge, and it replaces the older intent option inside If/Else conditions.

What each branch means

• Positive: satisfaction, agreement, enthusiasm.
• Negative: frustration, disappointment, disagreement.
• None: neutral wording, or text the AI cannot confidently classify.

Set it up

1. Open your workflow and click + where the check should happen.
2. Search for AI Intent Detection and add it.
3. Enter an Action Name such as "Check survey sentiment".
4. In Input Text, either type a fixed phrase or use the variable picker to insert dynamic text, for example the body of an incoming message.
5. Click Save. Three branches appear automatically.
6. Add steps to each branch.

Branch ideas

Example: post-delivery check-in

1. Trigger: an order is fulfilled.
2. Send an SMS asking how everything went.
3. Wait for a reply (use a Wait for Reply or a second workflow started by Customer Replied).
4. Pass the reply text into AI Intent Detection and handle each outcome.

Tips

• Feed it the customer's own words, not your outgoing message.
• Short answers like "ok" often land in None, so make that branch useful rather than empty.
• Sentiment is not consent. Never use a positive result as proof someone agreed to marketing.

Good to know: Always build the None branch; leaving it empty silently ends the contact's journey.