Agente de soporte al cliente GPT-5 RAG

Intermedio

Este es unSupport Chatbot, Multimodal AIflujo de automatización del dominio deautomatización que contiene 11 nodos.Utiliza principalmente nodos como Telegram, Agent, TelegramTrigger, LmChatOpenAi, EmbeddingsOpenAi. Construye un agente de soporte al cliente RAG usando GPT-5, Telegram y Pinecone

Requisitos previos
  • Bot Token de Telegram
  • Clave de API de OpenAI
  • Clave de API de Pinecone
Vista previa del flujo de trabajo
Visualización de las conexiones entre nodos, con soporte para zoom y panorámica
Exportar flujo de trabajo
Copie la siguiente configuración JSON en n8n para importar y usar este flujo de trabajo
{
  "meta": {
    "templateCredsSetupCompleted": true
  },
  "name": "GPT-5 RAG Customer Support Agent",
  "tags": [],
  "nodes": [
    {
      "name": "Disparador de Telegram",
      "type": "n8n-nodes-base.telegramTrigger",
      "position": [
        -340,
        20
      ],
      "parameters": {
        "updates": [
          "message"
        ],
        "additionalFields": {}
      },
      "typeVersion": 1.2,
      "id": "Disparador-de-Telegram-0"
    },
    {
      "name": "Modelo de chat OpenAI",
      "type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
      "position": [
        -100,
        240
      ],
      "parameters": {
        "model": {
          "__rl": true,
          "mode": "list",
          "value": "gpt-5"
        },
        "options": {}
      },
      "typeVersion": 1.2,
      "id": "Modelo-de-chat-OpenAI-1"
    },
    {
      "name": "Simple Memoria",
      "type": "@n8n/n8n-nodes-langchain.memoryBufferWindow",
      "position": [
        80,
        240
      ],
      "parameters": {
        "sessionKey": "={{ $json.message.chat.id }}",
        "sessionIdType": "customKey",
        "contextWindowLength": 15
      },
      "typeVersion": 1.3,
      "id": "Simple-Memoria-2"
    },
    {
      "name": "Pinecone Almacén de vectores",
      "type": "@n8n/n8n-nodes-langchain.vectorStorePinecone",
      "position": [
        200,
        220
      ],
      "parameters": {
        "mode": "retrieve-as-tool",
        "options": {},
        "pineconeIndex": {
          "__rl": true,
          "mode": "list",
          "value": "awm-n8n"
        },
        "toolDescription": "Customer FAQ Data & Policies"
      },
      "typeVersion": 1.3,
      "id": "Pinecone-Almac-n-de-vectores-3"
    },
    {
      "name": "Incrustaciones OpenAI",
      "type": "@n8n/n8n-nodes-langchain.embeddingsOpenAi",
      "position": [
        340,
        380
      ],
      "parameters": {
        "options": {}
      },
      "typeVersion": 1.2,
      "id": "Incrustaciones-OpenAI-4"
    },
    {
      "name": "Telegram",
      "type": "n8n-nodes-base.telegram",
      "position": [
        600,
        20
      ],
      "parameters": {
        "text": "={{ $json.output }}",
        "chatId": "={{ $('Telegram Trigger').item.json.message.chat.id }}",
        "additionalFields": {
          "appendAttribution": false
        }
      },
      "typeVersion": 1.2,
      "id": "Telegram-5"
    },
    {
      "name": "Nota adhesiva",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -480,
        -140
      ],
      "parameters": {
        "width": 320,
        "height": 700,
        "content": "Telegram Customer Support Trigger"
      },
      "typeVersion": 1,
      "id": "Nota-adhesiva-6"
    },
    {
      "name": "Nota adhesiva1",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -140,
        -140
      ],
      "parameters": {
        "color": 5,
        "width": 660,
        "height": 700,
        "content": "GPT-5 AI RAG Agent with Vector Database (Pinecone) Tool"
      },
      "typeVersion": 1,
      "id": "Nota-adhesiva1-7"
    },
    {
      "name": "GPT-5 Customer Support Agente",
      "type": "@n8n/n8n-nodes-langchain.agent",
      "position": [
        20,
        20
      ],
      "parameters": {
        "text": "={{ $json.message.text }}",
        "options": {},
        "promptType": "define"
      },
      "typeVersion": 2,
      "id": "GPT-5-Customer-Support-Agente-8"
    },
    {
      "name": "Nota adhesiva2",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        540,
        -140
      ],
      "parameters": {
        "color": 4,
        "width": 300,
        "height": 700,
        "content": "Telegram Response Output"
      },
      "typeVersion": 1,
      "id": "Nota-adhesiva2-9"
    },
    {
      "name": "Nota adhesiva3",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        -1120,
        -140
      ],
      "parameters": {
        "color": 6,
        "width": 600,
        "height": 960,
        "content": "🧠 RAG-Based Customer Support Agent (GPT-5 + Telegram)\nDescription:\n\nThis workflow builds a powerful Retrieval-Augmented Generation (RAG) Customer Support Agent that interacts with users directly through Telegram using the GPT-5 model. It combines real-time conversational capabilities with context-aware responses by leveraging vector search via Pinecone, making it ideal for automated, intelligent support systems.\n\nWatch Video Tutorial Build on Workflows Like These:\nhttps://www.youtube.com/@Automatewithmarc\n\n💬 Key Features:\n\nTelegram Integration: Listens to customer queries via the Telegram Trigger node and sends back intelligent responses in the same chat.\nGPT-5 Agent (LangChain): A powerful AI agent node orchestrates the conversation using OpenAI's GPT-5 model.\nContextual Memory: A Memory Buffer stores the last 15 interactions per user to provide more personalized and coherent multi-turn conversations.\n\nRAG with Pinecone: Integrates with Pinecone to fetch relevant answers from your “Customer FAQ” vector namespace, enabling grounded and accurate responses.\nEmbeddings Generation: Uses OpenAI’s Embeddings node to process and vectorize documents for retrieval.\nEnd-to-End AI Pipeline: Connects all components from input to output, providing seamless and intelligent customer support.\n\n🔧 Tech Stack:\n\nGPT-5 via OpenAI API\nPinecone vector store (namespace: Customer FAQ)\nTelegram Bot API\nLangChain agent, memory, and embedding tools\nn8n self-hosted or cloud instance\n\n📌 Ideal Use Cases:\n\nAutomated customer support for e-commerce, SaaS, or community support\nFAQ bots with up-to-date product or policy documents\nMultilingual support agents (customizable via GPT-5)\n\n🛠️ Setup Instructions:\n\nSet up your Telegram bot and insert credentials.\nAdd your OpenAI and Pinecone API keys.\nUpload or index your support documents into the Customer FAQ namespace on Pinecone.\nDeploy and test your Telegram bot."
      },
      "typeVersion": 1,
      "id": "Nota-adhesiva3-10"
    }
  ],
  "active": false,
  "pinData": {},
  "settings": {
    "executionOrder": "v1"
  },
  "versionId": "e5a5d5f7-38de-404d-a33d-4d341dec281b",
  "connections": {
    "Simple Memory": {
      "ai_memory": [
        [
          {
            "node": "GPT-5 Customer Support Agent",
            "type": "ai_memory",
            "index": 0
          }
        ]
      ]
    },
    "Telegram Trigger": {
      "main": [
        [
          {
            "node": "GPT-5 Customer Support Agent",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Embeddings OpenAI": {
      "ai_embedding": [
        [
          {
            "node": "Pinecone Vector Store",
            "type": "ai_embedding",
            "index": 0
          }
        ]
      ]
    },
    "OpenAI Chat Model": {
      "ai_languageModel": [
        [
          {
            "node": "GPT-5 Customer Support Agent",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "Pinecone Vector Store": {
      "ai_tool": [
        [
          {
            "node": "GPT-5 Customer Support Agent",
            "type": "ai_tool",
            "index": 0
          }
        ]
      ]
    },
    "GPT-5 Customer Support Agent": {
      "main": [
        [
          {
            "node": "Telegram-5",
            "type": "main",
            "index": 0
          }
        ]
      ]
    }
  }
}
Preguntas frecuentes

¿Cómo usar este flujo de trabajo?

Copie el código de configuración JSON de arriba, cree un nuevo flujo de trabajo en su instancia de n8n y seleccione "Importar desde JSON", pegue la configuración y luego modifique la configuración de credenciales según sea necesario.

¿En qué escenarios es adecuado este flujo de trabajo?

Intermedio - Chatbot de soporte, IA Multimodal

¿Es de pago?

Este flujo de trabajo es completamente gratuito, puede importarlo y usarlo directamente. Sin embargo, tenga en cuenta que los servicios de terceros utilizados en el flujo de trabajo (como la API de OpenAI) pueden requerir un pago por su cuenta.

Información del flujo de trabajo
Nivel de dificultad
Intermedio
Número de nodos11
Categoría2
Tipos de nodos8
Descripción de la dificultad

Adecuado para usuarios con experiencia intermedia, flujos de trabajo de complejidad media con 6-15 nodos

Autor
Automate With Marc

Automate With Marc

@marconi

Automating Start-Up and Business processes. Helping non-techies understand and leverage Agentic AI with easy to understand step-by-step tutorials. Check out my educational content: https://www.youtube.com/@Automatewithmarc

Enlaces externos
Ver en n8n.io

Compartir este flujo de trabajo

Categorías

Categorías: 34