This workflow follows the Gmail → Google Sheets recipe pattern — see all workflows that pair these two integrations.
The workflow JSON
Copy or download the full n8n JSON below. Paste it into a new n8n workflow, add your credentials, activate. Full import guide →
{
"name": "Finance RAG \u2014 Retrieval Agent",
"nodes": [
{
"id": "retrieval-001",
"name": "Telegram Trigger",
"type": "n8n-nodes-base.telegramTrigger",
"typeVersion": 1,
"position": [
240,
300
],
"parameters": {
"updates": [
"message"
],
"additionalFields": {}
},
"credentials": {
"telegramApi": {
"name": "<your credential>"
}
}
},
{
"id": "retrieval-002",
"name": "Embed Question",
"type": "n8n-nodes-base.httpRequest",
"typeVersion": 4,
"position": [
440,
300
],
"parameters": {
"method": "POST",
"url": "https://generativelanguage.googleapis.com/v1beta/models/text-embedding-004:embedContent",
"sendQuery": true,
"queryParameters": {
"parameters": [
{
"name": "key",
"value": "YOUR_GEMINI_API_KEY"
}
]
},
"sendBody": true,
"specifyBody": "json",
"jsonBody": "={{ JSON.stringify({ model: 'models/text-embedding-004', content: { parts: [{ text: $('Telegram Trigger').item.json.message.text }] } }) }}",
"options": {
"response": {
"response": {
"responseFormat": "json"
}
}
}
}
},
{
"id": "retrieval-003",
"name": "Query Pinecone",
"type": "n8n-nodes-base.httpRequest",
"typeVersion": 4,
"position": [
640,
300
],
"parameters": {
"method": "POST",
"url": "https://YOUR_PINECONE_INDEX_HOST/query",
"sendHeaders": true,
"headerParameters": {
"parameters": [
{
"name": "Api-Key",
"value": "YOUR_PINECONE_API_KEY"
},
{
"name": "Content-Type",
"value": "application/json"
}
]
},
"sendBody": true,
"specifyBody": "json",
"jsonBody": "={{ JSON.stringify({ vector: $json.embedding.values, topK: 3, includeMetadata: true, namespace: 'finance-kb' }) }}",
"options": {
"response": {
"response": {
"responseFormat": "json"
}
}
}
}
},
{
"id": "retrieval-004",
"name": "Build RAG Prompt",
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
840,
300
],
"parameters": {
"mode": "runOnceForEachItem",
"jsCode": "// Pull Pinecone matches\nconst matches = $input.item.json.matches || [];\n\n// Build context from retrieved chunks\nconst context = matches.length > 0\n ? matches.map((m, i) => `[Source ${i + 1}]\\n${m.metadata?.text || ''}\\n`).join('\\n')\n : 'No relevant policy sections found in the knowledge base.';\n\n// Get original question and user info from Telegram trigger\nconst telegramMsg = $('Telegram Trigger').item.json.message;\nconst questionText = telegramMsg.text;\nconst username = telegramMsg.from?.username\n ? `@${telegramMsg.from.username}`\n : (telegramMsg.from?.first_name || 'Unknown User');\nconst chatId = telegramMsg.chat.id;\n\nconst prompt = `You are a Finance Policy Assistant for a company. Your job is to answer employee questions accurately using ONLY the company's finance policy documentation provided below.\n\nRULES:\n- Answer using only the context provided. Do not use outside knowledge.\n- If the answer is not in the context, respond with exactly: \"I don't know. This query isn't covered in the finance policy documentation \u2014 I'm escalating it to the finance team.\"\n- Keep answers concise, clear, and professional.\n- If there are specific numbers (limits, thresholds, days), state them explicitly.\n\nFINANCE POLICY CONTEXT:\n${context}\n\nEMPLOYEE QUESTION: ${questionText}\n\nANSWER:`;\n\nreturn [{\n json: {\n prompt,\n questionText,\n username,\n chatId\n }\n}];"
}
},
{
"id": "retrieval-005",
"name": "Generate Answer",
"type": "n8n-nodes-base.httpRequest",
"typeVersion": 4,
"position": [
1040,
300
],
"parameters": {
"method": "POST",
"url": "https://generativelanguage.googleapis.com/v1beta/models/gemini-1.5-flash:generateContent",
"sendQuery": true,
"queryParameters": {
"parameters": [
{
"name": "key",
"value": "YOUR_GEMINI_API_KEY"
}
]
},
"sendBody": true,
"specifyBody": "json",
"jsonBody": "={{ JSON.stringify({ contents: [{ parts: [{ text: $json.prompt }] }], generationConfig: { temperature: 0.2, maxOutputTokens: 512 } }) }}",
"options": {
"response": {
"response": {
"responseFormat": "json"
}
}
}
}
},
{
"id": "retrieval-006",
"name": "Extract Answer",
"type": "n8n-nodes-base.code",
"typeVersion": 2,
"position": [
1240,
300
],
"parameters": {
"mode": "runOnceForEachItem",
"jsCode": "// Extract answer text from Gemini response\nconst candidates = $input.item.json.candidates || [];\nconst answerText = candidates[0]?.content?.parts?.[0]?.text\n || \"I don't know. I couldn't retrieve an answer \u2014 escalating to the finance team.\";\n\n// Detect low-confidence or out-of-scope answers\nconst uncertaintyPhrases = [\n \"i don't know\",\n \"i do not know\",\n \"not covered\",\n \"not in the context\",\n \"not found in\",\n \"i'm escalating\",\n \"escalating it to\",\n \"cannot answer\",\n \"no information\",\n \"not available in\"\n];\nconst lower = answerText.toLowerCase();\nconst isUncertain = uncertaintyPhrases.some(phrase => lower.includes(phrase));\n\n// Pull context from upstream nodes\nconst buildPromptData = $('Build RAG Prompt').item.json;\n\nreturn [{\n json: {\n answerText: answerText.trim(),\n isUncertain,\n questionText: buildPromptData.questionText,\n username: buildPromptData.username,\n chatId: buildPromptData.chatId,\n timestamp: new Date().toISOString()\n }\n}];"
}
},
{
"id": "retrieval-007",
"name": "Log to Google Sheets",
"type": "n8n-nodes-base.googleSheets",
"typeVersion": 4,
"position": [
1440,
300
],
"parameters": {
"operation": "append",
"documentId": {
"__rl": true,
"value": "YOUR_GOOGLE_SHEET_ID",
"mode": "id"
},
"sheetName": {
"__rl": true,
"value": "Audit Log",
"mode": "name"
},
"columns": {
"mappingMode": "defineBelow",
"value": {
"Timestamp": "={{ $json.timestamp }}",
"Username": "={{ $json.username }}",
"Question": "={{ $json.questionText }}",
"Answer": "={{ $json.answerText }}",
"Escalated": "={{ $json.isUncertain ? 'YES' : 'NO' }}"
},
"matchingColumns": [],
"schema": [
{
"id": "Timestamp",
"displayName": "Timestamp",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": true
},
{
"id": "Username",
"displayName": "Username",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": true
},
{
"id": "Question",
"displayName": "Question",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": false
},
{
"id": "Answer",
"displayName": "Answer",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": false
},
{
"id": "Escalated",
"displayName": "Escalated",
"required": false,
"defaultMatch": false,
"display": true,
"type": "string",
"canBeUsedToMatch": false
}
]
},
"options": {}
},
"credentials": {
"googleSheetsOAuth2Api": {
"name": "<your credential>"
}
}
},
{
"id": "retrieval-008",
"name": "Is Answer Uncertain?",
"type": "n8n-nodes-base.if",
"typeVersion": 2,
"position": [
1640,
300
],
"parameters": {
"conditions": {
"options": {
"caseSensitive": false,
"leftValue": "",
"typeValidation": "loose"
},
"conditions": [
{
"id": "condition-uncertain",
"leftValue": "={{ $json.isUncertain }}",
"rightValue": true,
"operator": {
"type": "boolean",
"operation": "equals"
}
}
],
"combinator": "and"
},
"options": {}
}
},
{
"id": "retrieval-009",
"name": "Gmail Escalation Alert",
"type": "n8n-nodes-base.gmail",
"typeVersion": 2,
"position": [
1840,
200
],
"parameters": {
"operation": "send",
"toList": "YOUR_FINANCE_TEAM_EMAIL@company.com",
"subject": "={{ '\u26a0\ufe0f Unresolved Finance Query from ' + $json.username }}",
"message": "={{ '<p><strong>Finance Policy Bot \u2014 Escalation Alert</strong></p><p>A query came in via Telegram that the bot could not confidently answer from the knowledge base.</p><p><strong>User:</strong> ' + $json.username + '</p><p><strong>Question:</strong> ' + $json.questionText + '</p><p><strong>Bot Response:</strong> ' + $json.answerText + '</p><p><strong>Time:</strong> ' + $json.timestamp + '</p><p>Please respond to this employee directly.</p>' }}",
"options": {
"isHtml": true
}
},
"credentials": {
"gmailOAuth2": {
"name": "<your credential>"
}
}
},
{
"id": "retrieval-010",
"name": "Telegram Reply \u2014 Escalated",
"type": "n8n-nodes-base.telegram",
"typeVersion": 1,
"position": [
2040,
200
],
"parameters": {
"operation": "sendMessage",
"chatId": "={{ $('Extract Answer').item.json.chatId }}",
"text": "={{ '\u26a0\ufe0f ' + $('Extract Answer').item.json.answerText + '\\n\\n_You will hear back from the finance team shortly._' }}",
"additionalFields": {
"parse_mode": "Markdown"
}
},
"credentials": {
"telegramApi": {
"name": "<your credential>"
}
}
},
{
"id": "retrieval-011",
"name": "Telegram Reply \u2014 Answer",
"type": "n8n-nodes-base.telegram",
"typeVersion": 1,
"position": [
1840,
420
],
"parameters": {
"operation": "sendMessage",
"chatId": "={{ $json.chatId }}",
"text": "={{ '\ud83d\udccb *Finance Policy Answer*\\n\\n' + $json.answerText }}",
"additionalFields": {
"parse_mode": "Markdown"
}
},
"credentials": {
"telegramApi": {
"name": "<your credential>"
}
}
}
],
"connections": {
"Telegram Trigger": {
"main": [
[
{
"node": "Embed Question",
"type": "main",
"index": 0
}
]
]
},
"Embed Question": {
"main": [
[
{
"node": "Query Pinecone",
"type": "main",
"index": 0
}
]
]
},
"Query Pinecone": {
"main": [
[
{
"node": "Build RAG Prompt",
"type": "main",
"index": 0
}
]
]
},
"Build RAG Prompt": {
"main": [
[
{
"node": "Generate Answer",
"type": "main",
"index": 0
}
]
]
},
"Generate Answer": {
"main": [
[
{
"node": "Extract Answer",
"type": "main",
"index": 0
}
]
]
},
"Extract Answer": {
"main": [
[
{
"node": "Log to Google Sheets",
"type": "main",
"index": 0
}
]
]
},
"Log to Google Sheets": {
"main": [
[
{
"node": "Is Answer Uncertain?",
"type": "main",
"index": 0
}
]
]
},
"Is Answer Uncertain?": {
"main": [
[
{
"node": "Gmail Escalation Alert",
"type": "main",
"index": 0
}
],
[
{
"node": "Telegram Reply \u2014 Answer",
"type": "main",
"index": 0
}
]
]
},
"Gmail Escalation Alert": {
"main": [
[
{
"node": "Telegram Reply \u2014 Escalated",
"type": "main",
"index": 0
}
]
]
}
},
"active": false,
"settings": {
"executionOrder": "v1"
}
}
Credentials you'll need
Each integration node will prompt for credentials when you import. We strip credential IDs before publishing — you'll add your own.
gmailOAuth2googleSheetsOAuth2ApitelegramApi
Pro
For the full experience including quality scoring and batch install features for each workflow upgrade to Pro
About this workflow
Finance RAG — Retrieval Agent. Uses telegramTrigger, httpRequest, googleSheets, gmail. Event-driven trigger; 11 nodes.
Source: https://github.com/toxicbishop/Customer-Behavior-Analysis-Power-BI/blob/c10a1608e32696b5b92127dd68fa4bd96e5c0608/n8n/retrieval_workflow.json — original creator credit. Request a take-down →
Related workflows
Workflows that share integrations, category, or trigger type with this one. All free to copy and import.
Arvifund - Supabase (Fixed v5). Uses httpRequest, telegram, googleSheets, telegramTrigger. Event-driven trigger; 91 nodes.
HTTP Request, Telegram, Google Sheets +9
Arvifund - Supabase. Uses httpRequest, telegram, googleSheets, telegramTrigger. Event-driven trigger; 90 nodes.
HTTP Request, Telegram, Google Sheets +8
Arvifund - Supabase (Fixed v2). Uses httpRequest, telegram, googleSheets, telegramTrigger. Event-driven trigger; 90 nodes.
HTTP Request, Telegram, Google Sheets +9
Arvifund - Supabase (Fixed v4). Uses httpRequest, telegram, googleSheets, telegramTrigger. Event-driven trigger; 90 nodes.
HTTP Request, Telegram, Google Sheets +9
Arvifund - Supabase (Fixed v3). Uses httpRequest, telegram, googleSheets, telegramTrigger. Event-driven trigger; 90 nodes.
HTTP Request, Telegram, Google Sheets +9