🎫 Automated customer feedback triage & follow-up with Gemini, Jotform, Gmail, Jira

⚑ 128 views · 🎫 Ticket Management & Triage

Description

AI-Powered Customer Feedback: Triage and Insight-Driven Chat

This n8n workflow creates a two-phase system for handling customer feedback received via a Jotform submission. The first agent quickly triages the issue, and the second agent engages in a persistent, conversational exchange over email to collect the information necessary for a resolution.

Phase 1: Triage and Initial Action (AI Agent)

This phase is triggered by a new submission on the Jotform. The goal is to immediately categorize the feedback and take the appropriate initial action.

  1. Jotform Trigger: The workflow starts instantly when a user submits your designated feedback form.
  2. AI Agent (Triage): This agent (powered by Google Gemini) is tasked with two primary jobs:
    • Sentiment Analysis and Response Drafting: It reads the feedback (q6_typeA6) and the user’s name (q3_name.first).
      • If Positive: It uses the Send a message in Gmail tool to send a concise, appreciative thank you note.
      • If Negative: It uses the Send a message in Gmail tool to send an initial, empathetic response acknowledging the issue and stating that a team member will follow up with questions. It also uses the Create an issue in Jira Software tool to log the bug or issue immediately.
    • Data Structuring: It uses the Structured Output Parser to extract key data points, most importantly the threadId of the initial email, which is crucial for the follow-up conversation agent.

Phase 2: Conversational Insight Gathering (AI Agent (Chat))

This phase takes over for all negative feedback, engaging the customer in a back-and-forth exchange to collect essential details required for the development or support team.

  1. Gmail Trigger: This node is set to poll for new, unread emails (which are expected to be replies from the customer).
  2. Simple Memory: This node is vital for the conversational aspect. It is configured to use the unique email threadId as its session key, allowing the AI Agent to remember the entire history of the conversation (previous questions asked and details provided) across multiple emails.
  3. AI Agent (Chat): This is the 2nd agent and the core of the conversational process.
    • Role: It acts as a dedicated feedback assistant.
    • Goal: Its instruction is to reply and ask for specific, missing information needed for the ticket, such as: what device they were using, if they know the steps to reproduce the issue, and to confirm that the team will send a free coupon for credits as a thank you for their help.
    • Tool: It uses the Reply to a message in Gmail tool to continue the conversation directly within the original email thread.
    • Resolution: The agent is trained to look for confirmation that all necessary information has been provided. Once it determines the issue details are complete, it will send a final thank you email and automatically use the Jira tool to summarize and update the existing Jira issue with the new insights, closing the loop on the data collection process.

Requirements

To implement and run this automated customer feedback workflow, the following accounts and credentials are required:

1. Automation Platform

2. Service Credentials

You must set up and connect the following credentials within your n8n instance:

3. External Configurations

πŸ”— Nodes Used

Jotform Trigger, Gmail Trigger, AI Agent, Simple Memory, Structured Output Parser, Google Gemini Chat Model

πŸ“₯ Import

Download workflow.json and import into n8n: Workflow menu β†’ Import from File

πŸ“– Importing guide Β· πŸ”‘ Credential setup