Redressing Posts
Redress means correcting the raw AI output of a social media check before you rely on the report.
Redress typically includes:
Removing false flags
Correcting behavioral attributes
Adjusting sentiment (when allowed for your use case)
Removing irrelevant web/news posts
Handling video/image media flags
Ferretly results are human-in-the-loop: AI retrieves and classifies public content; a person confirms accuracy.
Continue → Step 2
---
Step 2: Why redress matters
Ferretly automates profile discovery, post retrieval, behavior/keyword analysis, and web article relevancy scoring. Those steps remove repetitive work but cannot replace human judgment on false positives, context, or identity mistakes.
Redress ensures the Social Media Report and related outputs match what a trained reviewer would stand behind for your process.
Continue → Step 3
---
Step 3: Who performs redress
### Analyst Assistant
If your organization has Analyst Assistant, a Ferretly analyst redresses the subject. You usually receive the report after the subject is Redressed (green check badge).
### You review yourself
Without Analyst Assistant, your team filters flagged posts and completes the task guides in the next step.
### What Ferretly analysts do not do during redress
Verify intent behind keyword-flagged posts (subjective). Define keywords as narrowly as possible.
Change behavioral attribute definitions for one-off customer interpretations—feedback trains models and could harm overall accuracy.
Adjust individual post sentiment. For FCRA use cases, post sentiment is not included on reports.
Continue → Step 4
---
Step 4: Redress task map
Filter to flagged posts first (Find a Post), then:
Task | Guide |
|---|---|
--- | --- |
Remove or change behavioral flags | |
Frame/audio video flags | |
Correct sentiment | |
Remove web/news or redact images | |
Tag for later review |
After edits, refresh reports so downloads match the redressed data.
Continue → Step 5
---