Effective spam and automated-outreach filtering for the messaging inbox

User story

As a professional receiving dozens of messages per week, I want automated spam filtering that reliably identifies and quarantines templated sales pitches and bot-generated outreach, so that my inbox contains only genuine human communication.

Description

Message filtering must detect templated and AI-generated outreach at inbox level and route it to a separate folder rather than silently deleting it. Users must have controls to set thresholds, add senders to trusted or blocked lists, and filter by connection proximity. Filtering logic should be trainable by user feedback.

Acceptance criteria

1. Incoming messages are scanned and scored for spam likelihood before appearing in the main inbox
2. Messages above the spam threshold land in a "Filtered" folder — never silently deleted
3. User can adjust spam sensitivity: Strict / Balanced / Off
4. Connection request messages from non-connections default to "Requests" folder, not the main inbox
5. User can mark any message as spam or not-spam; feedback is used to improve their personal filter within 48 hours
6. Blocked senders cannot re-contact the user from new accounts (device/IP fingerprinting for repeat offenders)
7. Filter audit log shows how many messages were filtered in the last 30 days and why

Original request

The messaging inbox is flooded with spam and automated sales pitches. I need effective filtering that actually works.