Email Intent Classification: Automatically Categorize Incoming Email

The Overflowing Inbox Problem

Business inboxes receive dozens to hundreds of emails daily. Sales inquiries mix with support requests. Partnership proposals hide among newsletters. Important messages get buried.

Automated intent classification brings order to email chaos.

Email Intent Categories

Business Communication

• Sales Inquiry: Prospects asking about products/services
• Partnership Proposal: Collaboration and business development
• Meeting Request: Scheduling and calendar coordination
• Follow-Up: Continuing previous conversations
• Introduction: Networking and referrals

Customer Communication

• Support Request: Help needed with product/service
• Complaint: Dissatisfaction requiring attention
• Feedback: Suggestions and comments
• Billing Question: Payment and invoice inquiries
• Cancellation: Account or subscription termination

Operational

• Notification: Automated system alerts
• Newsletter: Marketing and content subscriptions
• Confirmation: Order, booking, registration confirmations
• Reminder: Scheduled follow-ups and deadlines
• Internal: Team and company communication

Building Your Email Classifier

Step 1: Define Your Workflow

What actions follow each email type?

Step 2: Train the Classifier

Describe your categories or upload sample emails. TinyModels generates diverse training examples matching email patterns.

Step 3: Integrate

Connect via API or use webhooks. New emails get classified instantly.

Step 4: Automate

Trigger actions based on classification:

• Route to folders/labels
• Create CRM records
• Generate tickets
• Send auto-responses
• Update spreadsheets

Integration Patterns

Gmail Integration

Use the Gmail API to process new messages. Fetch email content, classify with TinyModels, then apply Gmail labels and trigger downstream workflows based on the predicted intent.

Microsoft 365 / Outlook

Connect via Microsoft Graph API. Retrieve messages, classify them, and automatically move to folders based on classification results.

Zapier / Make Integration

1. Trigger: New email in Gmail/Outlook
2. Action: HTTP request to TinyModels API
3. Filter: Route based on classification result
4. Actions: CRM update, ticket creation, Slack notification

Use Cases by Industry

SaaS Companies

• Route trial inquiries to sales
• Escalate enterprise requests
• Auto-categorize support emails
• Identify churn signals in communication

Professional Services

• Categorize client requests by service type
• Route new business inquiries
• Flag urgent matters
• Sort administrative vs. billable communication

E-commerce

• Separate order inquiries from returns
• Route wholesale inquiries to sales
• Categorize product questions
• Identify influencer/partnership opportunities

Recruiting

• Classify applications by role/department
• Route referrals appropriately
• Identify passive candidate outreach
• Sort vendor vs. candidate communication

Measuring Success

Before Automation

• Time to triage: 30+ minutes/day
• Missed important emails: Weekly
• Response time variance: High
• Manual sorting errors: Common

After Automation

• Time to triage: 0 (automated)
• Missed important emails: Rare
• Response time: Consistent by category
• Classification accuracy: 90%+

Advanced Patterns

Priority Scoring

Combine intent with urgency. The classifier can return both the intent category and an urgency level, suggesting actions like "immediate_response" for high-priority complaints.

Sender Context

Enrich classification with sender data:

• Known customer → Higher priority
• VIP contact → Executive escalation
• Unknown sender → Standard routing

Thread Analysis

Classify entire conversations:

• Detect conversation drift (support → sales)
• Identify resolution status
• Track sentiment progression

Start Classifying

Your inbox shouldn't require constant manual attention. Build a classifier that understands your email types and routes messages automatically.

Describe your email categories. Get an API endpoint in minutes.