Supported Intents
Rypl reads every comment and message for what the person wants and gives it one intent label, plus a sentiment (Positive, Neutral or Negative). Intent detection is on every plan. The labels appear in the inbox, and you can filter by them and use them in reply rules.
The labels
| Intent | What it means | Examples |
|---|---|---|
| Question | Asks for information: price, availability, specs, how it works, where or how to get it | “How much?”, “Does it come in XL?”, “Do you ship to Spain?” |
| Purchase | Says outright they want to buy or order | “I’ll take two”, “How do I order? I want one” |
| Request | Asks you to do something: a return, refund, exchange, cancellation or order status check | “I want to return this”, “Where is my order #1042?” |
| Complaint | Unhappy with your product, service, price or staff, without asking for a specific action | “Arrived broken”, “Worst support ever” |
| Feedback | An opinion, suggestion or recommendation | “Love the new colours”, “Too expensive for me”, “You should make a smaller size” |
| Sharing | Tells a personal experience or circumstance without expecting a reply | “I’ve had mine for a year, still going strong” |
| Spam | Commercial spam, links to other shops, competitor mentions that send buyers elsewhere, self-promotion, bot text | “Cheaper at [another store]”, “DM me for a better deal” |
| User-to-user | Commenters talking to each other, often starting with @name | “@Anna look, this is the one I told you about” |
| Other | Greetings, simple emojis, anything unrelated | “Hi”, “🔥”, “first” |
When a comment both asks you to act and something else, Request wins: “This arrived broken, I want a refund” is a Request, not a Complaint.
How the labels are used
- Replies. Questions, purchase signals and requests are answered by default.
- Needs attention. Complaints and urgent comments wait for a person.
- Moderation. Spam is hidden automatically.
- Reply rules (Pro and Ultra). Use the label as a condition, for example “Type: Feedback, Sentiment: Positive → My text”.
- Inbox filters. Filter conversations by intent and sentiment.
What people write under ads
In RYPL’s Q3 2026 data, 40% of the comments RYPL’s AI analyzed on advertisers’ Facebook and Instagram ads and posts were questions, 38% showed buying intent (mostly questions about the product), and only 2% said outright that the person wanted to buy. Buying intent mostly arrives as a Question, not a Purchase.
More on rypl.is
- How to detect buying intent in ad comments: the questions that signal a buyer
- Ad comment statistics 2026: what people write under ads, RYPL data for Q3 2026
- Ad comment glossary: dark posts, hidden comments, Moderation Assist and other terms
| Back to Reference | Back to Documentation | Visit Rypl |