Skip to main content
There are several ways to create a persona, depending on how much data you’re starting from.

​
Generate From Filtered Signals

The most data-grounded path. Filter your signals to the audience you want to capture, then let the AI read that slice of your data and produce a full persona profile.
1

Filter your signals

Narrow signals down to the segment you want represented — for example, “trial users who mentioned pricing in the last 90 days” or “CS conversations about integrations.”
2

Preview the audience

Before generating, you’ll see the signal count, unique user count, company count, and a sample of the signals the persona will be built from. Adjust the filter if the sample doesn’t look right.
3

Add optional context

Add a workspace-level description (e.g., “We’re a B2B scheduling platform”) and an optional brief to steer the persona (e.g., “Focus on the IT admin view”).
4

Generate

The AI produces a complete profile — personal details, professional context, psychographics, and company context — with quotes tied back to specific signals.
5

Review and save

Adjust the name, pick a customer stage and product familiarity, optionally link a Knowledge page for extra context, and save to your persona library.

​
Generate From a Prompt

When you don’t have a pre-filtered slice in mind, start from a prompt and let the system recommend which persona types to build.
  • Enter a free-form prompt describing the kind of persona you want (e.g., “Skeptical mid-market buyer evaluating us against [competitor]”).
  • The AI looks at your org’s person type templates and recommends which ones to generate new personas from.
  • Pick the recommendations you want, refine if needed, and the system batch-creates the personas.

​
Generate in Bulk From Person Types

If you’ve set up custom person type templates (see Settings & Access), you can pick several at once and have the system generate one or more personas from each template. Useful when you want coverage across your whole audience quickly.

​
Create Manually

You can also build a persona by hand — fill in the name, profile details, customer stage, company context, and any quotes or behaviors you want to capture. Manual personas live alongside generated ones and can be edited the same way.

​
What’s in a Persona Profile

Each persona captures four layers of context:
LayerWhat’s Included
Personal profileName, job title, seniority, and any demographic detail the data supports (age, location, education, income)
Professional contextResponsibilities, goals, experience level, day-to-day schedule, and tools
PsychographicsMotivations, preferences, behaviors, needs, pain points, and representative quotes from the underlying signals
Company contextSize, industry, structure, budget, buying process, software stack, and strategic priorities
Generated personas cite the specific signals that shaped each claim. When you open a persona, you can trace any quote or pain point back to the real customer conversation it came from.

​
Persona Metadata

Every persona is also tagged with two fields you can edit anytime:
  • Customer stage — new prospect, trial, active customer, power user, or churned. Controls how the persona frames their relationship with your product.
  • Product familiarity — none, basic, intermediate, or advanced. Controls how much they assume in chats and tests.
Getting these right sharpens every downstream chat and test, so update them whenever a persona’s role shifts.

​
Editing and Organizing

From the persona library you can:
  • Open any persona to see their full profile, signal count, linked users, and linked companies
  • Edit any field — the AI-generated content is a starting point, not locked in
  • Link a persona to a Knowledge page so chats and tests can pull in extra context
  • Delete a persona when it’s no longer useful
Persona names are automatically de-duplicated at generation time, so two generated personas won’t end up with the same name. Manually created personas can share any name you pick.