Most Instagram audiences are treated as one undifferentiated inbox: everyone gets the same messages. This playbook changes that. You'll have the AI Agent collect data through natural conversation, sort contacts into the right segment automatically, and then target each segment with a campaign that actually matches what they want.
What you'll build
A system where the AI Agent learns key facts about each new contact, saves them to the profile, files the contact in the right folder based on what it learned, and feeds a targeted follow-up campaign for each segment.
Step 1: Decide what to collect and how to segment
Answer two questions: what do you need to know to send the right message, and what distinct groups do you want to target? A fitness coach might track experience level (beginner or advanced) and training preference (online or in-person), giving four segments with very different messaging. An e-commerce brand might split "buying for myself" from "buying as a gift". Write down 2 to 3 data points and 2 to 4 resulting segments before you build anything.
Step 2: Create a property for each data point
Create a contact property for each one, for example "Experience level" or "Training preference". For property types and setup, see Contact Profiles & Custom Properties.
Then go to AI Agent → User data collection and add each property. The agent now watches every conversation for these and saves values as it picks them up. See User Data Collection: Saving Contact Info Automatically.
Step 3: Create a folder for each segment
Create a folder for each segment you defined ("Beginners", "Online training", and so on). These are the buckets your contacts land in once the agent has done its work.
Step 4: File contacts with intent-based actions
In AI Agent → Actions and automations, create one intent-based action per segment. Set the intent (for example "identifies as a beginner looking for online training") and the action (add to the matching folder). The agent files contacts mid-conversation as soon as it has enough to make the call. See Intent-Based Actions: Links, Folders & Scenario Triggers.
Step 5: Build a campaign per segment
Create a campaign for each segment and set its audience to the matching folder or segment. Write the copy specifically for that audience: the beginner campaign can use different language and offers than the advanced one. That specificity is exactly why segmentation works. See Audience Selection & Segmentation for Campaigns.
Step 6: Send once the segments fill
Give the agent a week or two to populate the folders, then send. You can send immediately or schedule each campaign.
Variations
Use live segments instead of folders. Build a segment that filters on a property value (for example "Experience level is Beginner") so it updates automatically as new contacts arrive. See Segmentation & Filters: Build Smart Audiences.
Double-check before sending. Start each campaign scenario with a Scenario condition that confirms the contact is still in the right folder, so nobody gets a message meant for a segment they've moved out of.
Results to expect
Targeted campaigns built on real data consistently outperform one-size-fits-all blasts. As your folders fill and your copy gets more specific per segment, expect higher open and reply rates than a single broadcast to everyone.
🐾 Netsuke's Tips
The agent files contacts mid-conversation based on intent, sometimes before they've stated a preference outright. Check a few profiles after the first week to confirm the assignments look right.
Keep your property names tidy and consistent from the start. It keeps your segment filters readable later.
Don't over-segment. Two or three meaningful groups you actually write different copy for beats ten you don't.
What's next?
Once your campaigns are out, see which segment is converting. Head to Campaign Metrics & Performance Tracking for open rates, clicks, and conversions broken down by send.