Front and Slack: AI Sentiment Triage Template

Score the sentiment of every inbound Front conversation and alert a Slack channel the moment a customer sounds unhappy.

What It Builds

This Spojit template starts from a Webhook trigger that Front calls when a new conversation arrives. An Agent-mode Connector node reads the message and returns a structured score of positive, neutral, or negative. A Condition node then routes only the negative results to a Slack Connector node, which posts an alert to your support channel with the customer and a short reason. Positive and neutral messages pass through quietly, so the channel stays signal, not noise.

The Prompt

Paste this into Miraxa, the intelligent layer across your automation, and it builds the workflow, connecting the tools for you:

Build a workflow triggered by a Front webhook when a new conversation arrives. Add an Agent-mode Connector node that reads the message body and scores its sentiment as positive, neutral, or negative using a response schema with a sentiment field and a one-line reason field. Add a Condition node that checks if {{ sentiment.sentiment }} equals negative, and connect the true branch to a Slack Connector node that posts to the #support channel with the customer name, the conversation subject, and the reason.

Connectors Used

  • Webhook trigger - Front calls the workflow URL when a conversation is created, verified by a signing connection.
  • Front - pull conversation and contact detail with get-conversation and get-contact in Direct mode if the webhook payload is light.
  • Slack - post the alert with send-message to your chosen channel.

Customize It

Change #support to your escalation channel, and swap the routed value (for example, alert on both negative and neutral) by editing the Condition node in the properties panel. You can also widen the response schema to add an urgency field, or add a Front add-tag call so flagged conversations are labeled in the inbox.

Tips

  • Keep the sentiment node in Agent mode with a Response Schema so the output is always one of three values you can branch on cleanly.
  • Use a Direct-mode Slack node for the alert: it is deterministic, costs no AI credits, and posts the same format every time.
  • Reference the score with {{ sentiment.sentiment }} in the Condition node, and keep the reason short so the Slack message stays scannable.

Common Pitfalls

  • The Webhook trigger needs a signing connection to verify Front's calls; pick a scheme when you set it up or the trigger will reject requests.
  • If the webhook payload lacks the full message text, add a Front get-conversation step before scoring so the agent has the real body to judge.
  • Without a Response Schema the agent may return a sentence instead of a single label, which breaks the Condition match.

Related

Did this answer your question? Thanks for the feedback There was a problem submitting your feedback. Please try again later.