Skip to main content
Email Inbound automatically processes incoming emails, classifies them by type, and routes them to the appropriate workflow. The system uses AI to intelligently categorize messages and extract relevant information for automated handling.

​
What It Does

When an email arrives, the system:
  1. Classifies the email into one of three categories using AI
  2. Extracts relevant information based on the email type
  3. Creates tasks or processes quotes automatically
  4. Notifies admin teams of important messages
  5. Stores communication history in the database

​
Workflow Visualization

​
Email Categories

​
Service Message

Emails containing “Service Message” with a message ID format [numbers-alphanumeric] Processing:
  • Extracts message ID and cleans email body
  • Identifies company from sender’s email domain
  • Creates task with 3-day deadline
  • Stores in communication database
  • Sends notification to Admin and Sales teams

​
Send Quotation

Emails starting with “Follow up -” containing a quote ID format QUOTE-numbers-numbers Processing:
  • Extracts quote ID and service details (tipo_di_servizio, tipo_di_merce)
  • Triggers quote drafting workflow
  • Sends quotation to customer
  • Updates quote status to QUOTATION_SENT
  • Stores submission data in quotes database

​
Quote Request (Other)

General inquiries that may be service quote requests Processing:
  • AI detects if email is requesting a quotation
  • Extracts customer details (name, email, phone, language)
  • Creates new quote with status PENDING_GUEST_QUOTE
  • Sends automated reply in customer’s language asking for:
    • Type of service: Door-to-Terminal, Door-to-Door, Terminal-to-Door, Terminal-to-Terminal
    • Type of goods: Intermodal/Non-intermodal units, Bulk, Hazardous goods, NHM/IMO codes
  • Creates task for follow-up
  • Adds customer to CRM if new
Email Inbound Processing Flow

​
AI-Powered Features

​
Smart Classification

  • Uses AI models to analyze email subject and determine category
  • Handles variations like “Re:”, “Fwd:”, and parentheses in subjects

​
Message Cleaning

  • Removes email signatures and disclaimers
  • Extracts only the main message content for processing

​
Company Name Normalization

  • Matches sender’s domain against existing business database
  • Ignores common legal suffixes (Inc, LLC, Ltd, GmbH, etc.)
  • Returns official company name or creates new entry

​
Automatic Tag Generation

  • Creates descriptive 3-word tags for each email
  • Format: lowercase, hyphen-separated

​
Multilingual Responses

  • Detects email language automatically
  • Generates reply in customer’s native language
  • Supports any language, not limited to English/Italian

​
Technical Details

  • Trigger: Unauthenticated webhook (email forwarding endpoint)
  • Databases: datalake_main, ai-loop
  • Collections:
    • db_ai-loop_V3_Tasks - Email tasks with deadlines
    • db_ai-loop_quotes - Quote requests and status
    • db_ai-loop_V3_business - Company information
    • db_ai-loop_message_communication - Email history
    • db_ai-loop_System_Notifications - Admin alerts
    • users - Customer profiles
  • Notification Teams: Admin, Sales (configurable)