Comments and moderation

Comments under your posts are a public shop window: a quick, useful answer makes a good impression on everyone who reads it, not just on the person who wrote it. The agent's “Comments” tab governs this flow with two settings.

How it works in practice

1. 1

   A comment arrives

   Someone comments on one of your Instagram or Facebook posts. Comments you post from the business profile are ignored: the agent doesn't reply to itself.

2. 2

   The agent checks whether it is abusive

   If automatic moderation is on, the comment is classified first: clearly offensive/spam, or legitimate.

3. 3

   It hides or replies

   If it is abusive, the comment is hidden and the exchange ends there. Otherwise the agent writes a public reply under the comment.

4. 4

   Everything lands in the inbox

   Comment and reply appear in the “Comments” tab of your inbox, with a link to the original post.

Instructions for comment replies

The tab's first field governs how the agent writes public replies. It is separate from the direct-message instructions because public replies need different rules: more brevity, no personal data, an invitation to move to private. If you leave it empty, the agent reuses the direct-message instructions.

• “Keep public replies short, warm and on brand.”
• “Invite the person to message us privately for any detail or personal matter.”
• “Never share negotiated prices, order details or personal data in a public comment.”
• “Always thank people who leave a positive comment, with a different phrase each time.”

The “Generate” button is available here too, to get a draft you can adapt.

Comment moderation

The “Comment moderation” switch decides whether the agent may hide comments. When it is on the label next to it reads “Hides offensive comments”; when it is off it reads “Replies to all comments” and nothing is ever hidden.

• What gets hidden. Only obvious abuse: insults, hate speech, harassment, threats, sexual harassment, blatant spam and scams.
• What is NEVER hidden. Negative reviews, complaints, criticism, sarcasm, questions and neutral or positive comments. Hiding honest criticism would damage your reputation more than the comment itself.
• When in doubt, the comment stays. If the classification fails or is uncertain, the comment stays visible. The system is deliberately cautious.

Moderation rules (optional)

When moderation is on, a text field appears where you can describe in your own words what to hide and what to keep. These rules take precedence over the default behaviour, so you can make it stricter or more permissive for cases specific to your business.

Hide personal insults aimed at us or the staff
("you suck", "thieves") and any links to competitors or betting sites.

Always keep service complaints, questions about prices
and negative reviews written civilly visible.

From comment to private message

For requests that need personal data, a booking or a quote, the good practice is to invite the person to message you privately, where the agent can collect the information without exposing details in public. Set that behaviour in the comment-reply instructions.

Limits worth knowing

• Recent comments only. The agent handles comments that have just arrived, not your posts' historical archive: connecting the account will not trigger a burst of replies on old content.
• One reply per comment. Each comment gets at most one public reply. If you delete the conversation from the inbox, the comment will not be replied to again.
• It ignores its own comments. Comments published by the business profile are ignored, so the agent never starts a conversation with itself.
• Every action consumes an interaction. Both replying and hiding consume one of the interactions included in your plan. See “Plans, interactions and limits”.