By default, a trigger fires on all messages or all comments. AI-detected intent lets you be far more specific: instead of listing keywords, you describe what the person means in plain language and let AI decide whether it matches.
This works for both DM triggers and comment triggers.
⚠️ AI-detected intent needs AI features enabled on your account. If AI is off, the option is disabled.
What AI-detected intent does
When you pick AI-detected intent as your trigger filter, Inrō reads each incoming message or comment and checks whether the person's intent matches what you described. It's a natural-language match, so you don't need exact phrases or keyword lists.
A DM that says "hey how much does this cost?" and one that says "what's the price like?" both match the intent "asks about pricing," even though neither contains the word "pricing."
Set it up for a DM trigger
When configuring a Someone sends you a DM trigger, open the filter settings and select AI-detected intent.
A Describe message intent field appears. Type a plain-language description of what to catch, for example:
"wants to book a meeting"
"asks for a refund"
"asks about pricing or plans"
"wants to opt out of messages"
Inrō suggests relevant intents as clickable tags below the field, based on your account. Click one to add it quickly.
⚠️ Be as specific as you can. Broad intents like "asks a question" match almost everything; narrow ones like "asks about the cancellation policy" give precise, predictable triggers.
⚠️ AI-detected intent isn't always accurate. Use the Test intent detection field on the right of the panel to try a few real phrasings before you save.
Set it up for a comment trigger
When configuring a Someone comments on your post trigger, open the comment filter and select AI-detected intent. The setup is the same: describe the comment intent, use the suggested tags, and save.
Two related options sit in the comment trigger:
Exclude hateful and negative comments is on by default. AI filters out hate speech, harassment, and strongly negative comments before they trigger your scenario. Leave it on unless you have a reason to process every comment.
Allow comment replies to trigger automation is off by default, so only top-level comments fire the scenario. Turn it on if you want replies to comments to count too.
Intent versus condition: two different jobs
These two AI features are easy to mix up:
AI-detected intent (trigger filter) decides whether a scenario starts at all. It's evaluated once, on the incoming message, before the flow runs.
AI-detected condition (scenario action) decides how a running scenario branches. It's evaluated mid-flow, drawing on the full conversation and the contact's record.
You can use both in one scenario: intent to control which messages start it, and a condition to control what happens inside. The condition action is covered in AI Smart Actions in Scenarios.
🐾 Netsuke's Tips
Test before activating. Type a few ways real contacts might phrase the message into Test intent detection and confirm they're caught. Two minutes of testing prevents a week of misfires.
If you run several scenarios, avoid overlapping intents. If two both match "asks about pricing," both fire and the contact gets two replies. Keep intents distinct.
Keep Exclude hateful and negative comments on even when it isn't your main reason for using AI. It's a free layer of protection on high-visibility posts.
Pair an intent trigger with Hand over to AI Agent at the end of the flow: the trigger catches the right people, the scenario handles the structured part, and the agent handles the open conversation.
What's next?
Make the inside of your scenario smarter too. Read Adding AI to Your Scenarios for practical patterns, or AI Smart Actions in Scenarios for the full reference on every AI action. To review the trigger filters themselves, see Triggers: The Complete Reference.