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

# Set Up Your Project

> How to set up MentionPilot so your data actually makes sense

***

This is the part most people rush through.

And then a few days later they look at the dashboard and think,\
“this data doesn’t really help me.”

In most cases, it’s not the data, it’s the setup.

So take a few minutes here. It makes a big difference.

***

## Your Project Structure

<Card title="Foundation" icon="layer-group">
  Inside MentionPilot, your entire project is built around two things:

  the prompts you track

  the competitors you compare against
</Card>

Once those are in place, everything else (visibility, position, sentiment, trends) starts building automatically across platforms like AI Overviews, ChatGPT, Gemini, and Perplexity AI.

***

## Start with prompts (this is the foundation)

<Card title="Prompts" icon="message">
  Before anything else, you need to tell MentionPilot what conversations you care about.
</Card>

That’s what prompts are.

Not keywords. Not search terms. Actual questions people ask AI.

***

### A simple way to think about it

Open ChatGPT for a second and ask yourself:

“What would my customer type here before choosing a product like mine?”

That’s your prompt.

***

### Example

<Columns cols={2}>
  <Card title="Agency">
    “Which SEO agency should I hire for an eCommerce brand?”\
    “Best digital marketing agency for SaaS companies”
  </Card>

  <Card title="SaaS">
    “What’s the best CRM for small teams?”\
    “What tools help improve lead generation?”
  </Card>
</Columns>

***

### Why this matters more than it seems

AI doesn’t match keywords the way Google does.

It tries to understand:

* what the user wants
* what situation they’re in

So, a vague prompt gives vague results.

A specific prompt gives you much more useful data.

***

## What happens after you add a prompt

<Card title="Behind the scenes" icon="gear">
  Once you add a prompt in MentionPilot, a few things start happening in the background.
</Card>

We begin running that prompt across different AI platforms.

Then for each response, we track:

* whether your brand was mentioned (Visibility)
* where it appeared in the answer (Position)
* how it was described (Sentiment)
* when it was added (so you can track changes over time)

***

## Let’s break that down simply

<Card title="Metrics Overview" icon="chart-line">
  You don’t need to overthink the metrics, but you should understand what they mean when you see them.
</Card>

### Visibility

This is the easiest one.

Out of all the responses we collected for that prompt, how many included your brand?

If you’re showing up in 2 out of 10 responses, your visibility is low.

If you’re in 7 or 8, you’re doing well.

***

### Position

This is about placement.

If AI lists 5 tools and your brand is mentioned first, that’s strong.

If you’re mentioned last, the impact is much lower.

Over time, we show you the average position so you can see if you’re moving up or down.

***

### Sentiment

This is about how AI talks about you.

Are you being recommended positively?\
Just mentioned? Or described in a negative way?

This becomes important very quickly because users often trust the tone of AI responses.

***

## Each prompt has its own timeline

<Card title="Timeline" icon="clock">
  One thing a lot of users miss: Every prompt has its own history.
</Card>

You can go back and see:

* how your visibility changed
* when you started appearing more (or less)
* whether something you did had an impact

This is useful when you’re testing content, PR, or SEO changes.

***

<Card title="Next Step → Manage Your Setup" icon="arrow-right" href="/manage-setup">
  Continue to learn how to manage prompts and competitors effectively.
</Card>