Skip to main content
Set how your agent, the interviewer participants talk to, appears and speaks: the name, role, and avatar in the chat header, and the tone of voice, reveal policy, and form of address of every message. Pillow sets up the agent when it drafts your study. To change what the agent asks, see Write your interview guide; to give it context and rules, see Steer your agent with research notes.

​
Where the settings are

  • Draft study: in the Agent section of the study page in the Study Builder. Changes save automatically. You can also ask the builder chat, for example “Call the agent Maya and use a corporate tone”.
  • Live study: in Settings > Brief, in the Agent section. Your changes reach participants only after you click Save and publish. See Launch a study.
  • Ended study: the settings are read-only.

​
Set the name, role, and avatar

These fields make up the chat header participants see.
  • First name: the agent’s name, up to 40 characters. It also fills {{agent_name}} in your outreach message.
  • Gender: Female (the default), Male, or Non-binary. It sets how the agent refers to itself in languages with grammatical gender, and which portraits Generate offers.
  • Role: the line under the name, Researcher at {{org_public_name}} by default. {{org_public_name}} is replaced by the Public name in your organization settings. Pillow translates the role into each participant’s language.
  • Avatar: the picture next to the name. Company logo (the default) uses the logo from your organization settings. Custom photo takes a photo you upload or a portrait from Generate.
Show company logo, on by default, adds your logo to a Letter or Custom photo avatar.

​
Choose a tone of voice

Tone of voice sets how the agent words its messages: Casual (the default), Corporate, or Young and Friendly. It changes the wording, not what the agent asks. The tone does not rewrite your outreach message or the questions you already have. In the Study Builder, changing the tone offers Rewrite to redraft them in the new tone.

​
Reveal policy

Reveal policy, under Advanced settings in the Agent section, decides how the agent answers when a participant asks whether it is an AI. The default is Reveal AI Identity: the agent confirms briefly that it is an AI assistant helping your team, then returns to the interview. Every interview chat opens with this notice at the top of the conversation, translated into the participant’s language:
This interview is moderated by AI. Your responses are anonymous, so feel free to share your honest feedback.
The notice cannot be turned off or reworded, and it appears in preview and live interviews alike.

​
Use formal address

Formal address, also under Advanced settings, is Disabled by default. Set it to Enabled and the agent uses the formal form of address in languages that have one: vous in French or Sie in German. Translations of your outreach message and the agent’s role follow the same setting. In languages without the distinction, it changes nothing.

Interviews

How an interview runs for participants.