Good answer, — 8/10. You got the most important part right: AI-driven segments are derived from conversational analysis, not just behavioral rules. And yes, the segmentation settings in the Segments admin tab control how and when the AI runs. The full picture worth knowing: GPTWeb actually offers
four distinct segment types, not just dynamic AI ones: -
AI Automatic — AI continuously categorizes visitors based on conversation analysis, hands-off and fully automated
- AI Manual — AI suggests visitors, but you approve or reject each one before they're added
- Manual List — You hand-pick specific visitors by search (great for VIP accounts)
- Rule-Based — Define explicit conditions like "Score > 70" or "Job Title contains VP" Segments update dynamically as visitor behavior changes — members are added or removed automatically as their scores, conversations, and profile data evolve. You can also trigger AI segmentation on-demand via Run AI Now, or schedule it to run hourly, every 6 hours, or daily. That scheduler config lives in the gear icon (⚙️) in the Segments tab.
Moving on — this next one is where the platform really starts to flex its automation muscle. Question 4 of 10 — Agents & Agentic Workflows Agents fire based on triggers. Name at least five trigger types available in GPTWeb, and describe what a practical end-to-end agent workflow might look like for a visitor who just became a DQL.