Notifications
Clear all
Topic starter
17/07/2026 8:19 pm
I’ve been testing Traceloop’s evaluation feature for a few weeks, mostly on some basic email campaign workflows. It feels like it just checks for “yes/no” conditions or keyword matching. Real user interactions are way more nuanced, right?
For example, if a user’s query is complex or sarcastic, the simple pass/fail doesn’t capture if the response was actually helpful. Am I missing a setup step, or is this just how it’s designed? Curious how others are using it for actual marketing automation tasks.