n8n Multi-Channel Feedback Pipeline: AI Analysis → Jira Tickets → Monthly Reports

Unify scattered feedback from Telegram, Google Forms/Sheets, and Gmail into a normalized structure, then auto-classify, score, and route it into Jira tickets, Notion docs, and scheduled email reports

Who

Product managers, SaaS teams, and customer-support leads

You get

End-to-end automation from multi-channel intake to structured ticket creation, priority scoring, and monthly retrospectives with minimal manual triage

You need

n8n · jira · notion · google-sheets · gmail · telegram

Steps

1. Configure multi-channel triggers and credentials

   After importing the template into n8n, set up the three triggers: Telegram Trigger (listens for bot messages), Google Sheets Trigger (listens for new form/row additions with polling interval such as everyMinute), and Gmail Trigger (filters new emails by sender/label with polling interval such as everyHour). All three sources feed into the 'data normalizer' code node. Complete credential authorization for Google Sheets, Gmail, Notion, Jira Cloud, Telegram Bot, and OpenAI.

2. Normalize data and run AI enrichment

   The 'data normalizer' code node detects the source (telegram / gmail / google_forms) from payload structure and outputs a standard JSON with id, source, timestamp, user, content, and metadata. Then 'instant reply' sends a Telegram acknowledgment; 'user enrichment' maps the user tier (free/pro/enterprise) by email and attaches tier_weight. Finally the 'Message a model' OpenAI node classifies the feedback, extracts tags, scores sentiment/pain/impact/effort, and writes a summary with acceptance criteria.

   Execution prompt · Source text。OpenAI node 'Message a model' system message field

   Copy prompt

   You are a Product Manager analyzing user feedback.
   Your task is to categorize, tag, score and summarize feedback.
   
   Return ONLY valid JSON with this exact structure (no markdown, no backticks):
   {
     "category": "bug|feature_request|improvement|question",
     "tags": ["tag1", "tag2", "tag3"],
     "sentiment": "positive|neutral|negative",
     "pain_level": 1-10,
     "business_impact": 1-10,
     "implementation_effort": 1-10,
     "summary": "One sentence, max 80 chars",
     "description": "2-3 sentences for Jira ticket",
     "acceptance_criteria": ["criterion1", "criterion2", "criterion3"]
   }
   
   Rules:
   - Tags: max 3, specific and lowercase
   - Summary: actionable, clear, <80 chars
   - Scores: realistic based on content
   - Description: detailed enough for dev team
   - ONLY return the JSON, nothing else

   Execution prompt · Source text。OpenAI node 'Message a model' user message field

   Copy prompt

   =Analyze this user feedback:
   
   FEEDBACK: "{{$json.content.raw_text}}"
   USER TIER: {{$json.user.tier}}
   SOURCE: {{$json.source}}
   
   USER SELECTED CATEGORY: {{ $json.content.category_hint || 'none' }}
   USER URGENCY ASSESSMENT: {{ $json.content.urgency_hint || 'none' }}
   
   Return the JSON analysis.

3. Compute priority and create Jira ticket with docs and notifications

   The 'parse json' code node cleans and parses the AI response into ai_analysis JSON. The 'priority calculator' computes a RICE score from pain_level, business_impact, implementation_effort, and tier_weight, then maps internal priority P0-P3 to Jira priorities Highest-Low. 'Create an issue' opens a Jira ticket with summary, description, labels, and priority drawn from upstream nodes. 'post processing' assembles the Jira URL; if the source is Telegram, 'send result to user' sends a rich message with ticket link and scores. Simultaneously, 'log to analytics' appends a row to Google Sheets, 'end notification' sends an email confirmation, and 'Create a page' generates a linked Notion document.

   Execution prompt · Source text。Telegram node 'send result to user' text field

   Copy prompt

   =🎉 *Feedback treated !*
   
   🎫 *Ticket created  :* [{{ $json.jira.key }}]({{ $json.jira.url }})
   
   📊 *Analysis :*
   - Category : {{ $('priority calculator').item.json.ai_analysis.category }}
   - Priority : {{ $('priority calculator').item.json.scoring.priority }}
   - Score RICE :{{ $('priority calculator').item.json.scoring.rice_score }}
   
   🏷️ *Tags :* {{ $('priority calculator').item.json.ai_analysis.tags }}
   
   
   
   Thanks ! 🙏

4. Aggregate monthly data and generate the intelligence report

   '1 month Trigger' fires on a monthly schedule. 'query google sheet' reads the feedback log; 'aggregate monthly stats' computes totals, category breakdown, sentiment score, average RICE, P0/P1 rate, and top pain points/feature requests. Then 'Message a model1' uses AI to generate a structured monthly report (executive_summary, insights, trends, recommendations, risk_areas). 'parse response' extracts the JSON; 'monthly report' writes to Notion; 'montly report log' appends to Google Sheets; and 'Email reporting' sends a formatted HTML email to stakeholders.

   Execution prompt · Source text。OpenAI node 'Message a model1' system message field

   Copy prompt

   You are a Senior Product Manager analyzing monthly user feedback data.
   Generate strategic insights, identify trends, and provide actionable recommendations.
   
   Return ONLY valid JSON (no markdown):
   {
     "executive_summary": "2-3 sentence critical takeaway",
     "key_insights": [
       "Insight 1 (data-driven, actionable)",
       "Insight 2",
       "Insight 3",
       "Insight 4",
       "Insight 5"
     ],
     "trends": {
       "positive": ["Positive trend 1", "Positive trend 2"],
       "concerning": ["Concern 1", "Concern 2"]
     },
     "strategic_recommendations": [
       {
         "title": "Recommendation title",
         "rationale": "Why this matters (with data)",
         "priority": "High|Medium|Low",
         "estimated_impact": "Expected business impact"
       }
     ],
     "risk_areas": [
       {
         "area": "Risk description",
         "severity": "High|Medium|Low",
         "mitigation": "Suggested action"
       }
     ]
   }

   Execution prompt · Source text。OpenAI node 'Message a model1' user message field

   Copy prompt

   =Analyze this monthly feedback data for {{ $now.toFormat('LLLL yyyy') }}:
   
   OVERVIEW:
   - Total feedbacks: {{ $json.stats.total }}
   - Avg RICE Score: {{ $json.stats.avg_rice_score }}
   - P0/P1 rate: {{ $json.stats.p0_p1_rate }}
   - Sentiment score: {{ $json.stats.sentiment_score }}/10
   
   BY CATEGORY:
   {{ Object.entries($json.stats.by_category).map(([k,v]) => `- ${k}: ${v}`).join('\n') }}
   
   BY USER TIER:
   {{ Object.entries($json.stats.by_tier).map(([k,v]) => `- ${k}: ${v}`).join('\n') }}
   
   TOP PAIN POINTS (by RICE):
   {{ $json.stats.top_pain_points.slice(0, 5).map((p, i) => `${i+1}. ${p.summary} (RICE: ${p.rice}, Pain: ${p.pain})`).join('\n') }}
   
   TOP FEATURE REQUESTS:
   {{ $json.stats.top_feature_requests.slice(0, 5).map((f, i) => `${i+1}. ${f.summary} (RICE: ${f.rice})`).join('\n') }}
   
   Generate strategic insights and recommendations.

This is an official n8n workflow template. After importing it, complete the setup in this order: 1) Replace all placeholder IDs and URLs; 2) Authorize all external service credentials; 3) Verify the AI model selection in each OpenAI node (change the residual gpt-5 reference to gpt-4o or gpt-4o-mini); 4) Adjust tierMapping and RICE thresholds to match your business rules; 5) Run a test submission through each channel to confirm connectivity.