Personalized Email Outreach with AI, Telegram Bot, and Website Scraping

This n8n workflow automates the creation of personalized email outreach campaigns, leveraging AI for content generation, a Telegram bot for interaction, and website scraping for data collection. It streamlines the process of engaging with leads by generating tailored emails based on scraped website content and user input, and allows for human-in-the-loop approval via Telegram.

What it does

1. Triggers on Schedule or Telegram Input: The workflow can be initiated either by a scheduled interval (e.g., daily, weekly) or by a command received from a Telegram bot.
2. Scrapes Website Content: If triggered by a schedule, it fetches content from a specified URL (presumably from a list of leads or target websites).
3. Processes Scraped Data: The scraped content is transformed and prepared for AI analysis.
4. Generates Personalized Email Content with AI: An AI agent (using OpenAI's chat model and embeddings with a Postgres vector store for context) analyzes the scraped website content and a predefined prompt to generate a personalized email draft. It also utilizes Postgres chat memory for conversational context.
5. Sends Email Draft for Approval via Telegram: The generated email draft is sent to a Telegram chat for review and approval.
6. Waits for Telegram Approval: The workflow pauses, awaiting a "yes" or "no" response from the Telegram user.
7. Conditionally Sends Email:
   • If Approved: The personalized email is sent via an SMTP service.
   • If Rejected or No Response: The email is not sent, and the workflow might log the rejection or take alternative action (e.g., notify the user).
8. Handles Direct Telegram Commands: If the workflow is triggered by a Telegram command, it can respond directly to the user, potentially initiating a different branch of the workflow or providing information.
9. Manages Data: Uses a "Data table" node, likely for storing and managing lead information or campaign details.

Prerequisites/Requirements

• n8n Instance: A running n8n instance.
• Telegram Bot: A Telegram bot token and chat ID for sending and receiving messages.
• OpenAI API Key: An API key for OpenAI (for the AI Agent, Chat Model, and Embeddings).
• PostgreSQL Database: Access to a PostgreSQL database configured with the PGVector extension for the vector store and chat memory.
• SMTP Server: Credentials for an SMTP server to send emails.
• Website URLs: A list of target website URLs to scrape (if using the scheduled trigger).

Setup/Usage

1. Import the Workflow: Download the provided JSON and import it into your n8n instance.
2. Configure Credentials:
   • Telegram: Set up a Telegram credential with your bot token.
   • OpenAI: Set up an OpenAI credential with your API key.
   • PostgreSQL: Configure a PostgreSQL credential for connecting to your database (for PGVector and Chat Memory).
   • SMTP: Set up an Email Send (SMTP) credential with your email server details.
3. Update Node Parameters:
   • Telegram Trigger (Node 50): Ensure your Telegram chat ID is correctly configured.
   • HTTP Request (Node 19): Update the URL to the target website(s) you wish to scrape. This might involve dynamically pulling URLs from a data source.
   • AI Agent (Node 1119): Review the prompt and configuration for the AI agent to ensure it generates emails according to your requirements.
   • Postgres PGVector Store (Node 1278) and Postgres Chat Memory (Node 1267): Ensure these are correctly configured to connect to your PostgreSQL database and the relevant tables.
   • Send Email (Node 11): Configure the recipient, sender, subject, and body of the email, using expressions to dynamically insert AI-generated content.
   • Schedule Trigger (Node 839): Adjust the schedule as needed for automated runs.
4. Activate the Workflow: Once all configurations are complete, activate the workflow.

The workflow will then either run on its defined schedule, scraping websites and sending personalized email drafts for approval, or respond to commands sent to your Telegram bot.