✨ AI

Gmail AI Extractor

Automated Email-to-CRM Data Pipeline

Role: AI Automation Engineer

n8nGoogle Gemini 2.5 ProGmail APIGoogle OAuth 2.0Supabase (PostgreSQL)Monday.com APINext.jsVercel

Built an AI-powered Gmail automation platform that converts unstructured customer emails into structured CRM data automatically. The system securely connects users' Gmail accounts through OAuth 2.0, monitors CRM-related emails, extracts important customer information using Google Gemini 2.5 Pro, and automatically creates CRM records in Monday.com. All workflow executions and user activity are tracked through Supabase.

↗ Live App

What we built

Gmail Integration & Authentication

Built a Next.js frontend with secure Gmail OAuth 2.0 authentication.
Implemented Gmail account connection flow using Google authorization.
Managed access and refresh tokens securely through Supabase.
Added token validation and automatic refresh handling.

AI Email Processing Pipeline

Developed n8n workflows to automate email extraction and processing.
Integrated Gmail API to fetch CRM-related emails based on labels and subjects.
Parsed Gmail messages including: Sender information, Subject, Plain text content, HTML email body.
Used Google Gemini 2.5 Pro AI agent to extract structured CRM information: Customer name, Email, Phone number, Company, Message details, Lead priority, Lead source.
Generated clean JSON output for reliable CRM processing.

CRM Automation

Integrated Monday.com API to automatically create CRM records.
Mapped extracted AI data into CRM board fields.
Automated lead creation without manual copy-paste or data entry.

Workflow Monitoring & Reliability

Stored users and workflow execution history in Supabase PostgreSQL.
Added execution status tracking: Processing, Completed, Failed.
Implemented error handling and AI response validation.
Built reusable automation architecture supporting multiple Gmail users.

Impact

Automated the complete email-to-CRM workflow from inbox monitoring to lead creation.
Reduced manual email handling and data entry workload by approximately 70%.
Converted unstructured customer conversations into structured sales-ready CRM records.
Created a scalable AI automation framework that supports multiple users and inbox integrations.

← Back to Sales, CRM & Lead Automation