> ## Documentation Index
> Fetch the complete documentation index at: https://docs.trypillow.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Define your persona

> Give your agent context about who it's interviewing with persona attributes.

Persona attributes give your agent context about who it's interviewing, so the conversation stays relevant to that audience.

<Info>
  Persona is optional. Attributes give the agent context but don't restrict who can participate.
</Info>

## Adding attributes

1. In the study builder, click **Persona** in the right sidebar
2. Click **Add attribute**
3. Enter an attribute name (e.g., "Age") and value (e.g., "25-45")
4. Add more attributes as needed

## Editing and deleting

* **Edit**: Click on any attribute name or value to modify it inline
* **Delete**: Hover over an attribute and click the trash icon

### Duplicate key handling

If you add an attribute with a key that already exists (case-insensitive), the existing attribute is updated instead of creating a duplicate. For example, adding "age: 30-45" when "Age: 25-35" exists updates the existing entry.

## Formatting guide

Attributes are flexible. Use whatever format makes sense for your study:

| Key        | Example values                      |
| ---------- | ----------------------------------- |
| `age`      | `25-45`, `18+`, `65+`               |
| `location` | `US`, `California`, `San Francisco` |
| `plan`     | `free`, `premium`, `enterprise`     |
| `tenure`   | `<1 month`, `1+ years`              |
| `role`     | `manager`, `developer`, `designer`  |

<Info>
  You can use any attribute relevant to your study. These are just common examples.
</Info>

## How personas affect interviews

Your agent receives the persona as background on who it's interviewing. Use it for context the agent can't pick up from the conversation, like segment, plan, or tenure. It doesn't change the agent's tone or which questions it asks; those come from your **Agent** and **Interview Depth** settings.

## Best practices

Pick attributes that help answer your study questions, and use concrete values instead of vague labels:

| Less effective | More effective          |
| -------------- | ----------------------- |
| `young`        | `18-25`                 |
| `new user`     | `account age < 30 days` |
| `heavy user`   | `10+ sessions/week`     |

## What's next

<CardGroup cols={2}>
  <Card title="Customize outreach" icon="message-square" href="/guides/customize-outreach">
    Write engaging opening messages.
  </Card>

  <Card title="Multi-language study" icon="globe" href="/guides/multi-language">
    Reach users in their preferred language.
  </Card>
</CardGroup>