AI categorization for Front.com

Less junk mail,
more time for what matters

AI categorization automatically classifies incoming emails and archives junk messages. Focus on what matters — we handle the sorting.

Contact us → See how it works Back to overview
🇪🇺 GDPR compliant 🇫🇮 Server in Finland 🤖 Choice of AI provider 🔒 Encrypted traffic
3
Categories
24/7
Automatic operation
<2s
Processing time
Problem

Email takes too much time

Busy executives and entrepreneurs spend hours every week sorting messages. Important emails get lost in the noise.

📧

Spam and junk messages

Your inbox fills up with marketing emails, automated notifications, and other noise.

⏳

Manual sorting

Every message needs to be opened, evaluated, and moved to the right place. This repeats hundreds of times per week.

🚨

Urgent messages get buried

Important messages easily go unnoticed when your inbox is full of less urgent emails.

How it works

Three steps to a cleaner inbox

The service runs automatically in the background. You don't need to do anything.

1

Email arrives in Front

A new email triggers automatic processing. Your workday is not interrupted.

2

AI categorizes

AI evaluates the message and categorizes it: urgent, normal, or junk.

3

Automatic actions

Junk messages are archived, urgent ones are highlighted, and all get a clear label.

Features

Everything you need

AI categorization is designed for busy professionals who use Front for their personal email.

🤖

Smart categorization

AI analyzes message content and automatically categorizes it into three categories. Every message gets a clear label in Front.

🗃

Automatic archiving

Junk messages are automatically moved to archive when AI is confident enough. The threshold is adjustable.

🚨

Urgent messages highlighted

Urgent messages are highlighted with a special label and notification so they get immediate attention.

📊

Weekly summary

Every Monday you receive a summary of archived messages. See what was filtered and review if needed.

⚙️

Personal settings

The service adapts to your needs. Archive threshold, notifications, and categorization sensitivity are adjustable.

📈

Statistics and monitoring

Real-time statistics on categorizations, archiving, and server status. See how the service is performing.

Privacy & GDPR

Designed with privacy first

The service is built according to GDPR requirements. Personal data is processed minimally and persistent data is stored in the EU.

🇫🇮

Persistent data in EU

  • Server located in Finland
  • Front.com EU Data Pinning enabled
  • Choice of AI provider: OpenAI, MistralAI (EU), or local LLM (Finland)
  • All traffic HTTPS/TLS 1.2+ encrypted
🔒

Strong security

  • HMAC-SHA1 webhook signature validation
  • Timing-safe comparison against timing attacks
  • Separate API keys per customer
  • Secrets isolated from version control
✂️

Data minimization

  • Only text content is sent to AI (max 2000 characters)
  • Signatures and quoted threads are removed before AI processing
  • Clear PII is masked (`<EMAIL>`, `<PHONE>`, `<IBAN>`, `<ID>`)
  • Sender, recipient, and subject are removed
  • HTML formatting and attachments are removed
  • Messages are not stored on the server
📜

Contracts in order

  • OpenAI DPA + EU Standard Contractual Clauses (SCCs) or EU/local alternative
  • Front.com DPA (SOC 2, ISO 27001)
  • Lonteva DPA with customer
  • User opt-out option (exclude list)

What data AI receives — and what it doesn't

Original message

✗ Sender: name@company.com
✗ Recipient: support@business.com
✗ Subject: "Order #12345"
✗ Attachments: 3 pcs
✗ HTML formatting
✗ Timestamp
✓ Text content (max 2000 chars)
→
Filtering

Data sent to AI

✓ Only anonymized text content
No storage (Zero Data Retention)
No model training on customer data

Full transparency — Front's built-in monitoring tools

You don't have to take our word for it. Front admin users have access to built-in tools to independently monitor and verify the integration's behavior.

📋

Audit Log

  • Comprehensive log of all integration actions
  • See exactly what actions were taken and when
  • Comment additions, tags, and archiving with timestamps
  • Location: Front Settings → Company → Audit log
📈

API Usage Analytics

  • Real-time view of API call volumes and types
  • See which endpoints are being called
  • Rate limit monitoring
  • Location: Front Settings → Developers → Analytics
💡

Want to see the technical details? We have a separate technical privacy document covering architecture diagrams, contract structures, and detailed GDPR analysis. We're happy to present it.

FAQ

Questions and answers

The service connects to your Front account via webhook integration. Setup requires a Front API token and creating a webhook rule. We help with configuration.
Yes. Front settings include an exclude list. Messages from users on this list are not sent to webhook processing at all. Users can request to be added to the list.
No. Message content is processed in server memory and forgotten immediately after categorization. Only anonymous statistics are stored (e.g., "10 messages categorized as junk"). AI solution can be chosen based on customer needs: OpenAI, MistralAI (EU), or local LLM in Finland.
Each customer has their own confidence threshold. Only when AI confidence exceeds the threshold is a message automatically archived. The threshold can be raised or lowered as needed.
The weekly summary lists all archived messages, so incorrect archiving is quickly noticed. Messages can be restored from archive at any time via Front.
Summary

Everything ready

Area Status Description
AI categorization✓Urgent / normal / junk
AI provider choice✓OpenAI / MistralAI (EU) / local LLM (Finland)
Automatic archiving✓Adjustable threshold
Weekly summary✓Digest of archived messages
GDPR compliance✓Data minimization, PII masking, DPAs, SCCs
Webhook validation✓HMAC-SHA1 + timing-safe
Personal settings✓Customizable categorization and thresholds
Server in Finland✓Finnish jurisdiction
Encrypted traffic✓HTTPS/TLS 1.2+ on all connections

Interested?

We're happy to tell you more and show you how the service works in practice.

Contact us →