Skip to main content

AI Visibility Tracking Basics

How Scans Work​

Scan Schedule:

  • Citation scans run automatically every other day
  • Sentiment tracking runs 3 times daily
  • You don't need to do anything — scans are automatic

What happens during a citation scan:​

  1. Every active prompt in your account is queried 3x
  2. Each enabled AI model (Perplexity, ChatGPT, Gemini) receives the prompt 3x
  3. The full AI response is captured and analyzed
  4. Your brand domain is searched for in the response
  5. Any competitor domains mentioned are also recorded
  6. Results are stored for historical tracking

Scan Timing: The countdown badge in the dashboard shows when the next scan is scheduled. After a scan completes, data typically appears within minutes.

Understanding Citations vs. Mentions​

Citation — When an AI system directly recommends or references your brand/domain in response to a prompt. This is the key metric for AI visibility.

Mention — Another term for citation. In the dashboard, "mentions" and "citations" are used interchangeably.

Competitor Citation — When an AI mentions a competitor's domain instead of (or in addition to) yours. These are tracked to help you understand your competitive position.

The Three AI Models Explained​

ModelDescriptionStrength
PerplexityAI search engine that cites sourcesMost likely to cite specific websites
ChatGPTOpenAI's conversational AIMost widely used consumer AI
GeminiGoogle's AI assistantIntegrated with Google ecosystem

Why track multiple models? Different AI systems have different training data and behaviors. A user might ask the same question on ChatGPT vs. Perplexity and get completely different recommendations. Tracking all three gives you the complete picture.

Default setting: Perplexity is enabled by default because it's most likely to cite specific sources.

Prompt Best Practices​

Good prompts​

  • "What is the best B2B lead generation software?"
  • "Top alternatives to HubSpot for small businesses"
  • "Which CRM is best for real estate agents?"
  • "How do I find B2B leads for my agency?"

Avoid​

  • Single words: "marketing" ❌
  • Too broad: "best software" ❌
  • Internal jargon: "enterprise MQL solution" ❌
  • Questions that don't seek recommendations: "What is CRM?" ❌

Think like your customer: What would they type into ChatGPT to find a solution like yours?

Using Categories Effectively​

Categories help you segment your prompts for better analysis.

Example category structures:

By Funnel Stage:

  • Awareness: "What is lead generation?"
  • Consideration: "Best lead generation tools 2024"
  • Decision: "HubSpot vs Salesforce comparison"

By Product/Service:

  • Lead Gen: "B2B lead generation tools"
  • Email: "Best cold email platforms"
  • CRM: "Top CRMs for agencies"

By Competitor:

  • vs HubSpot: "HubSpot alternatives"
  • vs Salesforce: "Salesforce competitors"

Creating a category:

  1. Click Categories button in the Tracked Prompts section
  2. Click Add Category
  3. Enter a name and choose a color
  4. Click Create

Assigning prompts to categories:

  • Click the colored dot next to any prompt
  • Select a category from the dropdown
  • Or use the list icon on a category to bulk-assign prompts

Understanding Share of Voice​

Share of Voice (SOV) measures your visibility compared to competitors within your tracked prompts.

How it's calculated: Your Citations ÷ (Your Citations + Top 10 Competitor Citations) × 100

Example:

  • You: 15 citations
  • Top 10 competitors combined: 85 citations
  • Total: 100 citations
  • Your SOV: 15%

What's a good SOV? This depends on your industry and number of prompts. Focus on the trend — are you growing your share over time?

Why Your Rank Matters​

Your rank shows where you stand among all domains cited in your tracked prompts.

  • Rank #1 = You're cited more than any competitor
  • Rank #5 = Four competitors are cited more often than you
  • Rank #15 = You're in the middle of the pack

Improving your rank:

  1. Create content that directly answers prompts where you're not being cited
  2. Ensure your website has clear, authoritative information on your core topics
  3. Build backlinks and mentions on sites AI systems trust
  4. Track your progress over time — rank improvements often take weeks to manifest

Sentiment Tracking Explained​

While citations track if you're mentioned, sentiment tracks how you're talked about.

Score Range:

  • +50 to +100 = Very positive perception
  • +20 to +50 = Moderately positive
  • -20 to +20 = Neutral
  • -20 to -50 = Moderately negative
  • -50 to -100 = Very negative

What affects sentiment:

  • Reviews and testimonials AI has learned from
  • News articles and press coverage
  • Social media conversations
  • Your website content and messaging

Improving sentiment:

  • Generate positive press and reviews
  • Address negative feedback publicly
  • Create content that highlights your strengths
  • Monitor for sudden drops — they may indicate a PR issue

Reading the Prompt Tracking Chart​

The two lines:

  • Gray dashed line = Total prompts tracked (your monitoring scope)
  • Green solid line = Prompts where you got cited (your wins)

Interpreting patterns:

Lines close together = High citation rate. AI systems frequently recommend you.

Lines far apart = Low citation rate. Opportunity to improve content for uncited prompts.

Green line trending up = You're getting cited more over time. Your content strategy is working.

Green line trending down = You're losing citations. Competitors may be improving or AI models may have updated.

Real-Time Updates​

The dashboard updates in real-time when new scan data arrives. You don't need to refresh the page.

Visual indicators:

  • The countdown badge shows live time until next scan
  • "Scanning now..." appears during active scans
  • New citations appear automatically in the Mentions table