---
title: "Panels — query a whole audience at once | Minds"
canonical_url: "https://getminds.ai/guide/panels"
last_updated: "2026-05-19T11:20:03.872Z"
meta:
  description: "Groups of Minds queried in parallel. Build one from existing Minds or from a plain-language audience description. Compare segments side-by-side."
  "og:description": "Groups of Minds queried in parallel. Build one from existing Minds or from a plain-language audience description. Compare segments side-by-side."
  "og:title": "Panels — query a whole audience at once | Minds"
  "twitter:description": "Groups of Minds queried in parallel. Build one from existing Minds or from a plain-language audience description. Compare segments side-by-side."
  "twitter:title": "Panels — query a whole audience at once | Minds"
---

Minds Team

# **Panels — query a whole audience at once**

Groups of Minds queried in parallel. Build one from existing Minds or from a plain-language audience description. Compare segments side-by-side.

# Panels

A panel is a Group of Minds queried together. Instead of asking one person, you're asking your whole target audience at once — 8, 15, or 100 Minds responding in parallel, with answers auto-classified and aggregated.

Teams name them different things:

- **Marketing** — _customer panels_
- **Agencies** — _client insight panels_
- **Product** — _user panels_
- **Founders** — _expert panels / advisory boards_

The mechanic is the same: a representative sample of Minds, ready to query before a traditional brief is even finished.

## Building a Group — two paths

### Path 1 · Pick existing Minds

1. Select multiple Minds in the sidebar (multi-select).
2. Click **Create Group**.
3. Name it (_"Gen-Z Focus Group"_, _"Advisory Board"_).

### Path 2 · Describe your audience

When you don't have the right Minds yet, describe the segment and let the platform build it:

1. Click **New Group** (or mention a new group inline in the chat input).
2. Describe the audience — _"enterprise CTOs in San Francisco, SaaS, 500+ employees"_, _"working parents in Germany aged 30–45 with kids in school"_, _"B2B CMOs at Series-B companies"_.
3. Add any extra context that narrows the segment.
4. Hit **Generate**. A draft group appears with a representative sample of Minds — each with its own validated personality and profile.
5. Iterate: swap Minds in and out, add more, edit profiles, or accept and save.

This is the fastest way to go from _"I need to test this with my ICP"_ to _"I'm running the test."_

## Running a panel

Open the group in a chat. Ask your question, or drop in a stimulus — a landing page screenshot, a pitch deck, a competitor ad, a pricing page. Every Mind responds in parallel.

### How responses are classified

Each question is auto-classified as one of three types:

- **Scale** — numeric ratings (_"rate this from 1–10"_) → distribution + group averages
- **Categorical** — discrete options (_"yes / no / maybe"_, _"which of these three headlines?"_) → percentage breakdowns
- **Qualitative** — open-ended (_"what do you think about this concept?"_) → clustered themes

Follow-ups are understood in context. Ask _"do you like ice cream?"_ then _"which flavor?"_ — the system knows what "which flavor" refers to and reformulates it as a standalone question for each Mind.

## Two groups in one chat

This is where the insight is sharpest. Drop two (or more) groups into the same chat and ask one question — _"what's working on this landing page, what's not?"_

You get side-by-side views: _B2B marketing leaders_ vs. _mid-market SaaS CMOs_, for example. You see where they agree. You see where they pull apart. The contrast between segments is the research finding — not 12 people saying the same thing, but two calibrated audiences stress-testing your work at the same time.

## From panel to one-on-one

Sometimes one response catches your eye. Click the Mind's name and you're in a private one-on-one with just her, inside the same chat. Ask why. Challenge her logic. Show her a competitor. Ask what would change her mind. Then return to the panel view.

The panel gives you breadth. The pull-out gives you depth.

## Alignment

Every panel answer block has an **Alignment** dropdown in the header. Open it to see a per-group score (0–100%) labelled **High**, **Medium**, or **Low**.

The score is the average **response reliability** of the Minds in that group for that specific question. Each Mind's response is scored against its own persona definition — how well the answer matches who that Mind actually is — and we average those scores per group.

Read it as: _how on-character was this group's response_. High alignment means the segment answered consistently with its persona definitions; lower alignment is a flag to read the individual responses before you act on the aggregate.

Alignment is computed after the chart renders, so the dropdown shows a loading state on each row until the scores land.

This is a temporary metric. The proper group-alignment model — closeness to empirical research findings for that segment — is in development.

## Tips for good questions

- **Be specific** — _"rate the clarity of this tagline from 1–5"_ beats _"what do you think?"_
- **Mix persona types** — combine skeptics and enthusiasts in the same group for balanced feedback.
- **Iterate** — follow-up questions drill deeper into the interesting responses.
- **Run the same question across different groups** — the delta is the insight.

## Panel vs. chat

|  | Chat | Panel |
| --- | --- | --- |
| **Interaction** | back-and-forth conversation | question → structured group response |
| **Minds** | 1 or more, sequential | Group, all respond in parallel |
| **Output** | free-form messages | aggregated, classified, visualized |
| **Best for** | deep exploration with one persona | fast research across a segment |

---

_Panels turn your Minds into a research tool. Two groups, one question, two perspectives — that's the shortest path from question to insight._