Quick Start

:::summarybox achieve A brand set up and detected on every AI model Trakkr tracks A handful of prompts capturing what your customers actually ask A visibility baseline you can improve from - and trends to watch from day two

Before you start

You'll move faster with these to hand:

Your brand name the way customers say it - Nike, not Nike, Inc.
Your primary domain - nike.com
Two or three competitor names you want to compare against

No code, no integration, no waiting on data sources. If you don't have a Trakkr account yet, sign up first - the rest of this guide assumes you're logged in.

Step 1 - Add your brand · ~90 seconds {#add-brand}

Go to Settings → Brands and click Add Brand in the top right.

Three fields matter on day one. Trakkr grabs your logo from the website automatically.

Field	What to enter	Example
Brand name	How people refer to you	Nike
Website	Your primary domain	nike.com
Description	What you do, in a sentence	Athletic footwear and apparel

Step 2 - Pick your prompts · ~2 minutes {#pick-prompts}

Open Prompts from the sidebar and click Add prompt. (Or hit ⌘K and type add prompt - the global command palette works from anywhere.)

Prompts are the questions Trakkr will ask AI on your behalf - every model, every day. Write them the way a real person would ask, not the way you'd type into Google.

The bad ones aren't wrong - they're keyword-style queries. AI responds very differently to a fully-formed question than to a keyword. Match the shape of a real conversation.

Five to ten prompts is plenty to start. Cover three angles:

Discovery - "What's the best [category] for [use case]?"
Comparison - "How does [you] compare to [competitor]?"
Reputation - "Is [you] good for [specific need]?"

Don't try to be exhaustive - you'll add more once you've seen your baseline.

Step 3 - Run research · ~2 minutes of waiting {#run-research}

Open Research from the sidebar and click Run Research.

Trakkr asks every prompt to every AI model your plan covers (six on Free and Starter; eight on Growth and Scale, which add Perplexity and DeepSeek). Every response is analysed for brand mentions, position, and sentiment, and the results stream into your dashboard as they arrive.

This is the only step where you wait. Two minutes, give or take.

Step 4 - Read your baseline {#read-baseline}

Head to your Dashboard.

Four numbers tell most of the story. The first is your headline; the rest add nuance.

Metric	What it means	What's normal on day one
Visibility Score	How often you appear, weighted by where you rank	30-60% for established brands; 0-30% for newer ones
Presence	Percentage of prompts where you appear at all	40-70% is typical
Rank	Where you sit when AI lists brands (lower is better)	2-5 is healthy
Mentions	Raw count of brand mentions across every prompt × model	The number you'll watch grow over time

Below the hero, the model breakdown shows your performance on each AI individually. Scoring 80% on Claude and barely registering on Perplexity is completely normal - different training data, different retrieval, different audiences. The per-model split is where most diagnoses start.

For the formula behind every score on this page, see the Metrics.

You have data. Now what?

Three branches depending on what's on your mind right now.