How-to · 5 minutes per submission
How to handle a suspicious or fake campaign submission
Most submissions sail through automated verification. The ones that don't get flagged for review with a reason: account-history mismatch, screenshot inconsistency, URL freshness fail, ML fraud-model alert. The right response depends on which signal fired and how confident the system is.
Before you start
- A flagged submission in your dashboard
Steps
Open the submission and read the flag reason
The dashboard shows the specific signal: 'URL not freshly created', 'screenshot resolution mismatch', 'account history irregular', 'ML model flagged with 87% fraud probability'. Each tells you something different.
Verify against the public proof
Click through to the proof URL. If the post is live, public, and matches what was submitted, the flag is likely a false positive. If the post is gone, private, or mismatches, reject.
Reject with a clear reason if it's fake
Pick the most specific reject reason from the dropdown. The customer is notified with the reason and can resubmit if it was a misunderstanding. Don't write custom text in the reject reason — it's machine-readable.
Approve overrides if it's a false positive
If the proof is legitimate, click 'Approve override'. The verification engine learns from these — repeated overrides for similar patterns improve the model.
Escalate repeat offenders
Customers with multiple rejected submissions trigger an auto-suspension. You don't need to do anything — the platform handles it. If you see a pattern across customers (e.g. all from the same account network), report it to support.