Skip to main content
A panel is a group of personas you can test against together. Instead of running a test against one persona at a time, you build a panel that looks like your target audience and hit it with one prompt.

What You Can Put in a Panel

Each panel member is either:
  • A single persona — one respondent per member
  • A replica set — multiple AI-generated variations of one base persona (up to 100)
Replica sets are what let a panel feel like a real audience. Instead of one “VP of Ops” weighing in, you can have 25 slightly different VPs of Ops react to the same prompt and see the distribution.

Creating a Panel

1

Name the panel

Give it a name and an optional description describing what it represents (e.g., “Mid-market buyers — ops and RevOps”).
2

Add members

Add as many members as you want. For each, pick either a single persona or a replica set.
3

Configure replica sets

For each replica set, choose the base persona, set the replica count (1–100), and pick a variation strength: low, medium, or high.
4

Save

The system auto-generates the synthetic respondents behind each replica set so the panel is ready to test immediately.

Variation Strength

Variation strength controls how much each replica differs from its base persona:
SettingWhat It DoesWhen to Use
LowSmall variations in wording, tone, and priorities while keeping the persona nearly identicalWhen you want a tight audience — e.g., a specific role at a specific segment — and just need multiple data points
MediumModerate variation across preferences, behaviors, and context while staying in the same segmentDefault for most tests — balances realism with coherence
HighBroader variation — different companies, slightly different roles, different motivations — while staying in the intent of the base personaWhen you want to stress-test an idea across a wider slice of a segment
Start with medium for a new panel. If all replicas end up saying the same thing, bump to high. If reactions feel all over the place, drop to low.

Managing a Panel

From the panel detail page you can:
  • See total member count and total respondent count (base personas + every replica)
  • Reorder members to control how they appear in reports
  • Edit or delete the panel
  • Launch a test run with the entire panel as the audience

When to Use Panels vs. Single Personas

  • Use a single persona for focused work — one-on-one chats, message rehearsals, or quick gut checks.
  • Use a panel when you want signal across a group — concept tests, A/B choice tests, or distribution-style results (“how does sentiment vary across 30 respondents in this segment?”).
A panel doesn’t lock you into testing everyone at once. You can still run a test against a subset of persona IDs even when they’re part of a panel.