MaveraMavera Docs
Operator's Guide
Checking…
Use or share this pageSend the page itself, print it, or open its complete context in your AI workspace.

Building audiences

The audience is the product — here's how to make good ones.

2 min readDocs build Updated Jul 26, 2026
Interactive product walkthrough

Build an audience inside the platform

Use the sidebar, audience builder, grounding source, facets, and launch-to-chat flow.

Simulation onlyNothing is saved or charged
1
Step 1 of 7Open Audience Builder

Choose Audience Builder from the sidebar.

Demo workspace / Mavera workspace
Local simulation
Mavera workspace

Good afternoon

Choose a surface from the sidebar to begin.

Demo workspace0 actions completed
WorkspaceDemo companySafe local simulation
Current surfaceaudience BuildFollow the highlighted action
Recent workspace activityStatus
Audience comparisonReady to review
Customer interview evidenceGrounded
Campaign pre-flightDraft
  1. 1
    Open Audience BuilderChoose Audience Builder from the sidebar.
  2. 2
    Start a new audienceCreate a new audience instead of editing an existing one.
  3. 3
    Describe real peopleWrite a specific audience with role, context, geography, and buying behavior.
  4. 4
    Choose the grounding sourceSelect whether this audience starts from a description, uploaded data, or a prior study.
  5. 5
    Create the audienceBuild the local demo audience and inspect its generated facets.
  6. 6
    Open the audienceOpen the finished audience and verify the name, facets, and grounding state.
  7. 7
    Talk to itLaunch a conversation from the finished audience card.

Everything on Mavera runs against an audience: a synthetic population built to think, talk, and decide like a specific group of real people.

Three ways to build one

1. A sentence (fastest)

Open Audience Builder and describe them like you'd brief a recruiter:

"Mid-30s suburban parents in the Southeast who hike on weekends and own at least one dog."

Auto mode picks the pathway, builds the population, and shows you the composition before you use it.

2. Your own data (sharpest)

The current Ingest page extracts completed research studies into source-quoted dossiers. It does not yet turn raw interviews, reviews, support transcripts, or CRM exports into grounded audiences. Build an audience from a clear category definition, then keep any human-study evidence separate from modeled responses.

3. From a past study

Any study's respondents can become a reusable audience ("build audience from study") — useful when a screener question found you an interesting subgroup.

What makes a good audience description

  • Who they are (age, place, role): "Raleigh-area consumers who bank with a national institution"
  • What they're doing (behavior): "…and have little awareness of local alternatives"
  • What moves them (motivation): "…but reconsider when a life event or service frustration hits"

That's the pattern the best-performing audiences use — identity, behavior, trigger.

Using audiences across workspaces

Audiences belong to a workspace. From My Research, the ⧉ button on any audience copies it into another workspace you belong to — the copy records where it came from.

Chatting WITH an audience

In the Mave composer there's an 👥 Audience picker. Select one or more audiences and your chat becomes a conversation with them — Mave answers from their worldview, and honestly declines questions their profile can't support. All of Mave's abilities stay available: mid-chat you can say "now run a proper study on this" and she will.

Honesty note

An audience built from a one-sentence description knows only what that description implies plus category knowledge. It will say "that's not something my profile covers" rather than invent. If you need deeper answers, feed it deeper data via Ingest.

Workshop: competitor teardown

A cited, structured read on a competitor — then war-game the response with your audience.

Talking to your audience

Chat, focus groups, interviews, and debates — when and how.

On this page

Three ways to build one1. A sentence (fastest)2. Your own data (sharpest)3. From a past studyWhat makes a good audience descriptionUsing audiences across workspacesChatting WITH an audience