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

# Prompts

> Define the natural-language questions Hikoo runs against AI assistants to test your visibility

A **prompt** is a natural-language question that Hikoo runs against AI assistants to see whether your brand gets cited. Your prompts are the tracked query set behind every citation, so writing good ones is the foundation of the whole dashboard.

## The prompts page

The Prompts page shows your prompts as a list, with stat cards across the top for quick health: **All**, **Winning**, **At risk**, **Unclaimed**, and **Average visibility**.

Each prompt row summarizes how you are doing on that question:

* **Visibility score** and **citation rate**
* The **platforms** that cite you on it
* The **competitors** cited alongside you
* The **markets** it is tracked in
* **Sentiment** of the answers
* A **status** of winning, at risk, losing, or unclaimed

## Add a prompt

You can add prompts three ways.

<Tabs>
  <Tab title="Manually">
    Write a single prompt yourself. Use this when you know the exact question you want to track.
  </Tab>

  <Tab title="Generate with AI">
    Give Hikoo a set of keywords and let it draft prompts for you. Review and keep the ones that fit.
  </Tab>

  <Tab title="Import CSV">
    Drop a CSV into the import dropzone. Hikoo shows a preview, auto-creates any missing topics, and highlights row-level errors before you commit the import.
  </Tab>
</Tabs>

## Classify prompts

Classification helps you slice your prompts by how buyers actually ask. Each prompt can carry:

* A **search intent**: informational, comparative, transactional, or navigational.
* A **journey stage**: discovery, consideration, decision, or retention.
* A **source keyword** it came from.

These values can be AI-generated, so you can classify a large set without tagging each one by hand.

## Organize and act in bulk

Group prompts by **topic**, by **journey stage**, or by none, then search and filter to find what you need. Select prompts to run bulk actions: **select all**, **delete**, **tag**, **retarget market**, **move to topic**, or **export**.

## Prompt volume

**Prompt volume** estimates how much AI search demand a prompt has, so you can prioritize the questions people actually ask. It is available on some plans.

## Prompts beyond your plan

Your plan sets how many prompt slots each website gets, and a usage bar shows your current use. If you go over that limit, the extra prompts are paused under **Beyond your plan** and can be restored. Adding more slots reactivates them.

<Warning>
  Paused prompts stop collecting new citations until you restore them.
</Warning>

## Prompt detail

Open a prompt to see how it has performed:

* **Visibility over the last 30 days**
* Your **best rank** and **coverage**
* A **per-platform breakdown**
* The **latest answers** AI assistants gave
* The **citations** and **competitor citations** tied to the prompt

<Tip>
  Write prompts the way real users ask AI assistants. Natural, specific questions match how people actually query, which makes your tracking reflect real demand.
</Tip>

## Related

<Columns cols={2}>
  <Card title="Topics" icon="folder" href="/guides/topics">
    Group prompts into themes that drive filtering everywhere.
  </Card>

  <Card title="Citations" icon="quote" href="/guides/citations">
    The answers your prompts produce, matched and ranked.
  </Card>

  <Card title="Search queries" icon="search" href="/guides/search-queries">
    What AI assistants actually search while answering your prompts.
  </Card>
</Columns>