AI Moderator Scopes

The AI moderator adapts its behavior based on scope — from hands-off structured tasks to deeply adaptive cross-session learning. Choose the right level of AI involvement for your research question.

01 The hierarchy
Study
Session 1Participant A
T1
T2
T3
Session 2Participant B
T1
T2
T3
Session 3Participant C
T1
T2
T3
02 Four scopes
Default for items

None

No AI involvement. Structured tasks only. Participants move through the flow at their own pace without dynamic interrogation.

Works with: comparison, rating, categorization
Observer mode

Text Monitor

Mic on, AI listens and may intervene with text prompts. Captures think-aloud observations and nudges for clarity without introducing voice bias.

Works with: URL browsing, topic discussions, free text
Most popular

Session

AI moderates within one session, adapting in real time. It asks about low ratings, follows up on free-text responses, and probes friction during URL tests.

Each session is independent — no cross-session memory
“I found the checkout process a bit confusing...”
“Which step specifically felt confusing?”
Advanced

Study

AI learns across ALL sessions. Identifies themes from early sessions, probes gaps in later ones. Use coverage reports to track what's been explored.

Best for 5+ session discovery research
03 Scope × Task type matrix
ScopeComparisonRatingCategorizationFree TextURL TestTopic Interview
None
—
Text Monitor
Session
—
Study———
04 How to choose
Decision 01

Need structured item evaluation?

Use None
Decision 02

Want think-aloud observations?

Use Text Monitor
Decision 03

Want adaptive interviews?

Use Session
Decision 04

Running 5+ sessions, want cross-session learning?

Use Study

Ready to try AI moderation?

$curl -fsSL https://candor.sh | bash