Skip to main content
Persona Chat lets you talk directly with a persona. The AI stays grounded in the persona’s profile, the signals that shaped it, and any Knowledge pages you linked — so the conversation reflects what that kind of customer would actually say.

​
Two Chat Modes

Talk at mode

Roleplay as the persona in first person. The AI responds as “I would…” — useful for rehearsing a sales pitch, testing a cold email, or simulating a support conversation.

Talk with mode

Analyze the persona in third person. The AI responds as an analyst — “they would likely…”, “their main concern is…” — useful for research prep, positioning work, or persona onboarding for your team.
Both modes use the persona’s full profile and underlying signal data. They just answer from different vantage points.

​
Starting a Chat

  1. Open any persona from your library
  2. Start a new chat on their detail page
  3. Pick Talk at or Talk with when you start
  4. Send messages — responses stream back in real time
You can have multiple chats with the same persona — each is saved separately with its own history.

​
What a Persona Knows

When you chat, the AI has access to:
  • The persona’s full profile — personal, professional, psychographics, and company context
  • The signals that were used to generate the persona (and any new signals matching their segment)
  • Any Knowledge page you linked to the persona for extra context
Personas can cite specific signal IDs when they back up a claim, so you can click through to the underlying customer data.

​
Good Uses for Chat

  • “How would you react to this email subject line?”
  • “What’s the first objection you’d raise to this pitch?”
  • “Rewrite this value prop in your own words.”
Chat is great for fast iteration on one idea. When you need signal across a group — a panel of dozens of respondents — run a Persona Test instead.