AI-powered email automation for business: Summarize & respond with RAG

This workflow is ideal for businesses looking to automate their email responses, especially for handling inquiries about company information. It leverages AI to ensure accurate and professional communication.

How It Works

1. Email Trigger:

   • The workflow starts with the Email Trigger (IMAP) node, which monitors an email inbox for new messages. When a new email arrives, it triggers the workflow.

2. Email Preprocessing:

   • The Markdown node converts the email's HTML content into plain text for easier processing by the AI models.

3. Email Summarization:

   • The Email Summarization Chain node uses an AI model (DeepSeek R1) to generate a concise summary of the email. The summary is limited to 100 words and is written in Italian.

4. Email Classification:

   • The Email Classifier node categorizes the email into predefined categories (e.g., "Company info request"). If the email does not fit any category, it is classified as "other".

5. Email Response Generation:

   • The Write email node uses an AI model (OpenAI) to draft a professional response to the email. The response is based on the email content and is limited to 100 words.
   • The Review email node uses another AI model (DeepSeek) to review and format the drafted response. It ensures the response is professional and formatted in HTML (e.g., using <br>, <b>, <i>, <p> tags where necessary).

6. Email Sending:

   • The Send Email node sends the reviewed and formatted response back to the original sender.

7. Vector Database Integration:

   • The Qdrant Vector Store node retrieves relevant information from a vector database (Qdrant) to assist in generating accurate responses. This is particularly useful for emails classified as "Company info request".
   • The Embeddings OpenAI node generates embeddings for the email content, which are used to query the vector database.

8. Document Vectorization:

   • The workflow includes steps to create and refresh a Qdrant collection (Create collection and Refresh collection nodes).
   • Documents from Google Drive are downloaded (Get folder and Download Files nodes), processed into embeddings (Embeddings OpenAI1 node), and stored in the Qdrant vector store (Qdrant Vector Store1 node).

Set Up Steps

1. Configure Email Trigger:

   • Set up the Email Trigger (IMAP) node with the appropriate IMAP credentials to monitor the email inbox.

2. Set Up AI Models:

   • Configure the DeepSeek R1, OpenAI, and DeepSeek nodes with the appropriate API credentials for text summarization, response generation, and review.

3. Set Up Email Classification:

   • Define the categories in the Email Classifier node (e.g., "Company info request", "Other").
   • Ensure the OpenAI 4-o-mini node is configured to assist in classification.

4. Set Up Vector Database:

   • Configure the Qdrant Vector Store and Qdrant Vector Store1 nodes with the appropriate Qdrant API credentials and collection details.
   • Set up the Embeddings OpenAI and Embeddings OpenAI1 nodes to generate embeddings for the email content and documents.

5. Set Up Document Processing:

   • Configure the Get folder and Download Files nodes to access and download documents from Google Drive.
   • Use the Token Splitter and Default Data Loader nodes to process and split the documents into manageable chunks for vectorization.

6. Set Up Email Sending:

   • Configure the Send Email node with the appropriate SMTP credentials to send responses.

7. Test the Workflow:

   • Trigger the workflow manually using the When clicking ‘Test workflow’ node to ensure all steps execute correctly.
   • Verify that emails are summarized, classified, and responded to accurately.

8. Activate the Workflow:

   • Once tested, activate the workflow to automate the process of handling incoming emails.

Key Features

• Automated Email Handling: Automatically processes incoming emails, summarizes them, and generates professional responses.
• AI-Powered Classification: Uses AI to classify emails into relevant categories for targeted responses.
• Vector Database Integration: Retrieves relevant information from a vector database to enhance response accuracy.
• Document Vectorization: Processes and stores documents from Google Drive in a vector database for quick retrieval.
• Professional Email Formatting: Ensures responses are professionally formatted and concise.

Need help customizing?

Contact me for consulting and support or add me on Linkedin.