Intent Detection

Intent detection reads every comment and message for what the person wants: information, to buy, an action from you, or to complain. Rypl labels each one with an intent and a sentiment, on every plan, with nothing to set up.

What you see

Each comment in the inbox shows its intent label when it carries one:

Question · Purchase · Request · Complaint · Feedback · Sharing · Spam · User-to-user

Comments labelled Other (greetings, a lone emoji) don’t get a badge. Every label is defined, with examples, in Supported Intents.

Sentiment is Positive, Neutral or Negative.

What it drives

Intent	What Rypl does by default
Question, Purchase, Request	The agent answers
Complaint	Goes to Needs attention for a person
Spam	Hidden automatically
Feedback (positive)	Answered with a thank-you reply if you turn those on (Pro and Ultra)
Sharing, User-to-user, Other	Left alone unless a reply rule says otherwise

Using intent in reply rules (Pro and Ultra)

Reply rules use intent (shown as Type) and sentiment as conditions:

Type: Feedback · Sentiment: Positive → My text: “Thanks for the love! 💜”
Type: User-to-user → Don’t reply
Type: Complaint → AI replies, with instructions to apologise and give a next step

Using intent to improve your ads

Many questions about the same thing (sizes, shipping) mean the ad or landing page leaves it out.
Many complaints on one ad often point to a promise the product doesn’t keep.
Mostly spam on one ad suggests turning on the Strict Spam & Media Filter for that channel.

Next Steps

Languages.