A Guide to Content Moderation and Why Your Messages Get Flagged

Bulk messaging platforms are an obvious target for abuse - spam, harassment, hate speech at scale - and SimpliSend runs every outgoing message through an automated content check before it sends, as a POPIA-aligned safeguard rather than a filter on ordinary business communication. Understanding how it works makes the rare case where a legitimate message gets flagged much less confusing.

What it's checking for

The check looks for hate speech, abuse, racism, and spam-like patterns - categories chosen because they represent genuine harm and platform misuse, not because they touch on ordinary sales or reminder language. A normal marketing message, payment reminder, or appointment confirmation essentially never comes close to tripping it.

Two severities, two outcomes

A flagged message is scored LOW or HIGH severity. LOW severity lets the message send as normal but queues it for human review under your account's moderation section - useful for borderline cases that are probably fine but worth a second look. HIGH severity blocks the send outright before it ever reaches the network; that message shows a BLOCKED status and never gets delivered.

How the classification actually happens

Message content is checked by an AI classifier trained to recognise hate speech, abuse, and racism in context, which handles nuance - sarcasm, reclaimed language, context-dependent phrases - far better than a simple keyword list could. If that classifier is ever unavailable, a heuristic fallback checks for generic threat and profanity patterns instead; it's deliberately not a slur list, and deliberately more conservative, since a fallback should err toward flagging for review rather than silently waving something through.

If your message gets flagged

• Check your account's moderation section for the specific flag and category
• A LOW-severity flag on a legitimate message doesn't need action - it still sent, and review is informational
• A HIGH-severity block means the message didn't send - review the wording and rephrase before resending
• If a message seems wrongly flagged, it's still visible for review rather than silently discarded, so nothing about it is hidden

For most businesses this system runs invisibly in the background, doing exactly its job of catching the rare genuinely harmful message without ever touching a normal campaign, reminder, or transactional notice.