Knowledge Base

Everything you need to understand AI visibility, interpret your metrics, and get the most out of Lighthouse.

What Is AI Visibility?

5 min readGetting Started

AI Visibility is the measure of how often and how accurately your brand appears in AI-generated responses when people ask questions related to your industry, products, or services.

Think of it like this: when someone asks ChatGPT "What are the best project management tools?", some brands will be mentioned by name and others won't. The brands that consistently appear across AI platforms have high AI visibility. The ones that don't — regardless of how well-known they are — are invisible in this channel.

Why It Matters

Consumer behavior is shifting fast. People increasingly use AI assistants not just for quick answers, but for product research, service comparisons, and purchase decisions. When an AI assistant recommends your competitor but not you, that's a lost opportunity you might never see in your analytics.

Unlike traditional search where you can track rankings and clicks, AI conversations are largely invisible to brands. You can't see the question that was asked, you can't see the response that was given, and you can't see whether you were mentioned or not. Lighthouse makes this visible.

How Lighthouse Measures It

Lighthouse takes a systematic approach:

  • We crawl your site to understand your content, key topics, and brand messaging
  • We generate relevant prompts — the kinds of questions your potential customers would actually ask AI
  • We query 10+ AI platforms with those prompts — ChatGPT, Claude, Gemini, Perplexity, DeepSeek, Qwen, and more
  • We analyze every response to check if your brand was mentioned, how it was described, what sentiment was used, and what sources were cited

The result is a comprehensive scorecard showing your visibility percentage, sentiment, citations, and competitive position across the entire AI ecosystem.

Key takeaway: AI Visibility is to AI assistants what search ranking is to Google. It's the measure of whether your brand shows up when it matters most — and Lighthouse is the tool that makes it visible and actionable.

Suggestions or corrections

The Lighthouse Diagnostic: Seen → Said → Sourced

4 min readGetting Started

Every AI visibility problem comes down to one of three failures. The Lighthouse diagnostic gives each one a name, a question, and a fix — so instead of staring at a dashboard wondering where to start, you ask three questions in order.

1. Seen — does AI mention you at all?

The question: when a buyer asks AI about your category, are you in the answer?

Where to look: your AI Visibility Score and the AI Response Testing tab. Every prompt where competitors appear and you don't is a Seen failure.

The fix: coverage. AI can't recommend what it never encounters — publish content that answers the exact questions buyers ask, in the categories where you're absent. Your Keywords & Queries tab shows which questions have demand.

2. Said — what does AI say when it mentions you?

The question: when you ARE in the answer, is the framing accurate and positive — or hedged, outdated, and wrong?

Where to look: your Sentiment score, the Insight Brief, and the verbatim responses in AI Response Testing. Being mentioned with caveats ("however, some users report…") can be worse than not being mentioned.

The fix: narrative control. Correct stale facts at the sources AI reads, publish first-party answers to the objections AI keeps raising, and monitor sentiment swings with weekly scheduled analyses.

3. Sourced — does AI build its answer from YOUR pages?

The question: of all the sources AI cites when discussing your space, how many are yours?

Where to look: the Citations tab. If AI describes you entirely through third parties — review sites, Reddit threads, competitor comparisons — you don't control the story even when it's positive.

The fix: citability. Structure your content so AI can lift from it directly (your GEO Score measures exactly this), and chase the citation gaps — sources that cite competitors but not you are your PR hit list.

Why the order matters

Fixing Said before Seen wastes effort — there's no narrative to control if you're not in the answer. Fixing Sourced before Said amplifies a story you haven't shaped. Work the beats in order: get Seen, then control what's Said, then become the Source.

You'll find the diagnostic throughout Lighthouse: the Seen / Said / Sourced strip under your Insight Brief shows all three beats with your live numbers, and the results tour walks them in sequence.

Running Your First Analysis

3 min readGetting Started

Getting your first AI visibility report takes less than 3 minutes. Here's exactly what to do.

Step 1: Enter Your Website URL

Go to the Lighthouse app and type your website's domain into the URL field. You don't need the "https://" part — just something like yourcompany.com.

Step 2: Choose Your Settings (Optional)

Click "Advanced Options" if you want to customize your analysis. Most people can skip this for their first run, but here are the options:

  • Analysis Mode: Auto Discovery (default) finds pages automatically. Tracked Pages analyzes the same pages every time. Custom URLs lets you specify exact pages.
  • Custom Topics: Add specific topics you care about — like "sustainable fashion" or "enterprise security"
  • Competitors: Add competitor URLs to compare your visibility against theirs

Step 3: Click "Analyze"

Hit the Analyze button and Lighthouse will start working. The process takes 1-3 minutes depending on how many pages are being analyzed. You'll see a progress indicator (and a mini-game to pass the time).

Step 4: Explore Your Results

When the analysis is complete, you'll see your dashboard with five key scores at the top: AI Visibility, Sentiment, GEO Score, Alignment, and Citations. Click any tab to dive deeper into specific areas.

Pro tip: Your first analysis establishes your baseline. Run a second analysis in a week to start tracking changes over time. If you're on Pro or Agency, you can schedule this automatically.

Suggestions or corrections

Reading Your Dashboard

4 min readGetting Started

Your Lighthouse dashboard is organized into five score cards at the top and multiple deep-dive tabs below. Here's how to read it all.

The Five Score Cards

AI Visibility
The percentage of AI platforms that mention your brand when asked relevant questions. Higher is better. Above 50% is strong; below 25% needs attention.
Sentiment
How positively or negatively AI describes your brand, on a 0-100 scale. 55+ is positive, 45-54 is neutral, below 45 is negative.
GEO Score
How "citable" your content is to AI models. This measures whether your pages have the structure, depth, and authority that AI tends to reference.
Alignment
How well your actual content matches the topics AI associates with your brand. High alignment means your content reinforces what AI already says about you.
Citations
The total number of times AI platforms linked back to a source when discussing your brand's topics. More citations mean more traceable traffic potential.

The Tabs

Below the score cards, you'll find specialized views:

  • Overview — Executive summary, key metrics, and the "What's New" AI narrative
  • Pages — Page-by-page content scores and optimization opportunities
  • GEO Score — Deep dive into AI citability and topic alignment
  • Citations — Where AI gets its information when discussing your topics
  • Keywords — Keyword visibility research
  • Commerce — Shopping query monitoring (how AI recommends products)
  • Site Traffic — AI bot crawler tracking
  • Influencers — Influencer analysis
  • Visualization — Interactive topic map showing how AI clusters your content
  • Prompts — Custom prompt testing across all AI platforms
  • Strategy — Recommendations and action items
  • Focus Group — Synthetic audience testing
Pro tip: Hover over any metric with a ? icon to see its definition, benchmarks, and recommended strategy — right in the dashboard.

Suggestions or corrections

AI Visibility Score

4 min readCore Metrics

Your AI Visibility Score is the single most important metric in Lighthouse. It answers the question: "When people ask AI about my industry, does my brand come up?"

How It's Calculated

Lighthouse asks relevant prompts across 10+ AI platforms. For each prompt, we check whether the AI's response mentions your brand by name. Your visibility score is the percentage of responses where your brand appeared.

For example, if Lighthouse tested 20 prompts across 10 platforms and your brand appeared in 85 out of 200 total responses, your visibility score would be 42.5%.

What "Good" Looks Like

70%+ (Excellent)
Your brand is well-established in AI's knowledge base. AI consistently recommends you across multiple platforms and prompt types.
40-69% (Good)
You appear in many responses but there are gaps. Look at which platforms and which topics are missing you.
20-39% (Moderate)
You have presence but it's inconsistent. There's significant room to improve through content optimization.
Below 20% (Low)
AI rarely mentions your brand. This is common for newer brands or those in competitive industries. Focus on building authoritative content around your core topics.

How to Improve It

  • Create comprehensive, authoritative content around the topics AI associates with your industry
  • Get mentioned in publications and sources that AI models frequently cite
  • Use structured data, FAQ sections, and clear brand mentions on your key pages
  • Monitor which specific prompts you're missing and create content that addresses those exact queries
Think of it this way: AI Visibility is like "share of voice" in traditional PR, but for AI-generated conversations. The higher your percentage, the more often AI is recommending you to potential customers.

Suggestions or corrections

Sentiment Score

3 min readCore Metrics

Your Sentiment Score measures how positively or negatively AI platforms describe your brand when they do mention you. Being visible is important — but being described well is equally critical.

How It's Calculated

When AI mentions your brand, Lighthouse analyzes the language and tone of the surrounding context. Each response gets a sentiment score from 0 (very negative) to 100 (very positive), and your overall score is the average across all responses.

Reading Your Score

55-100 (Positive)
AI describes your brand with favorable language — words like "leading," "popular," "well-regarded," "innovative." This is where you want to be.
46-54 (Neutral)
AI mentions you factually but without strong opinion. Not bad, but there's an opportunity to strengthen your narrative.
0-45 (Negative)
AI uses unfavorable framing — words like "controversial," "criticized," "limited," "expensive compared to." This needs immediate attention.

What Influences AI Sentiment

  • Reviews and press coverage: AI models absorb sentiment from product reviews, news articles, and discussions they were trained on
  • Comparison language: If your brand is consistently positioned as "the budget option" or "less feature-rich than X," that shapes AI sentiment
  • Your own content: The way you describe yourself on your website can influence how AI frames you to others
Watch out: A brand can have high visibility but negative sentiment — meaning AI mentions you often but not favorably. Monitor both metrics together.

Suggestions or corrections

GEO Score (Generative Engine Optimization)

5 min readCore Metrics

Your GEO Score measures how likely your website content is to be cited and referenced by AI models. Think of it as "SEO for AI" — while traditional SEO helps you rank in search results, GEO helps you get mentioned in AI conversations.

What Makes Content AI-Citable?

AI models tend to cite content that is:

  • Authoritative — Backed by data, expert quotes, or original research
  • Structured — Uses clear headings, lists, tables, and FAQ formats that AI can easily parse
  • Comprehensive — Covers a topic thoroughly rather than superficially
  • Unique — Provides original insights rather than rehashing what others have said
  • Current — Published or updated recently with up-to-date information

How Lighthouse Scores It

Lighthouse evaluates each page on your site across these dimensions and assigns a GEO score from 0-100. Pages are categorized as Strong (70+), Moderate (40-69), or Weak (below 40).

Quick win: Pages with FAQ sections, data tables, and expert quotes consistently score higher. If you have blog posts that are mostly opinion with no supporting data, those are your biggest GEO opportunities.

Suggestions or corrections

Topic Alignment

3 min readCore Metrics

Topic Alignment measures how well your website content matches the topics and themes that AI associates with your brand and industry.

Why Alignment Matters

If AI thinks your industry is about "sustainable packaging" and "eco-friendly materials" but your website talks mostly about "competitive pricing" and "fast delivery," there's a disconnect. High alignment means your content reinforces the narrative AI is already building about your space.

How to Read It

Lighthouse breaks alignment into three levels:

  • Strong pages (70+): Your content closely matches what AI associates with your topics
  • Moderate pages (40-69): Partial match — some topics are covered but gaps exist
  • Weak pages (below 40): Significant disconnect between your content and AI's expectations

The Topic-by-Topic Breakdown in the GEO Score tab shows you exactly which topics have strong alignment and which need work.

Strategy: Use the alignment data to prioritize content creation. If AI expects to see content about "industry regulations" and you don't have any, that's a clear content gap to fill.

Suggestions or corrections

Citations

3 min readCore Metrics

Citations tell you where AI goes to back up its claims. When an AI assistant says "According to [source]..." — that's a citation. Your Citations metric shows how many of those references point to your domain versus competitors.

Why Citations Matter

Citations are the closest thing to a "backlink" in the AI world. When AI cites your website, it's telling users that your content is trustworthy enough to reference. This drives actual traffic and builds credibility in a way that simple mentions don't.

What to Look For

  • Your domain citations: How many times AI cited pages on your website
  • Competitor citations: Who else AI cites when discussing your industry
  • Source domains: Which publications and sites AI trusts most in your space
Strategy: If you see that AI frequently cites a particular publication in your industry, getting featured or quoted in that publication could boost both your citations and your visibility.

Suggestions or corrections

Custom Prompt Testing

3 min readTools & Features

Custom Prompt Testing lets you ask any question across all AI platforms simultaneously and see how each one responds. This is your research lab for understanding AI behavior in real time.

How to Use It

Navigate to the Prompts tab in your dashboard. Type any question — "What is the best CRM for small businesses?" or "What companies are leading in sustainable fashion?" — and click search. Lighthouse will query every AI platform and show you the responses side by side.

What to Look For

  • Consistency: Do all platforms mention the same brands? If some mention you and others don't, you can target the gaps.
  • Framing: How does each platform describe your brand differently? One might call you "affordable" while another says "premium."
  • Competitor presence: Which competitors appear most frequently and in which platforms?
Pro tip: Test prompts that match your actual customers' language. Don't just test "best [product category]" — test the way real people ask: "I need a [product] that does X and Y, what should I get?"

Suggestions or corrections

AI Bot Tracking

3 min readTools & Features

AI Bot Tracking shows you when AI crawlers visit your website — which pages they read, how often they return, and which AI companies are actively indexing your content.

How It Works

Install a lightweight tracking script on your website (one line of code). Once installed, Lighthouse detects and logs visits from known AI bots including GPTBot (OpenAI), ClaudeBot (Anthropic), Google-Extended, PerplexityBot, and others.

What You'll See

  • Bot visit frequency: How often each AI crawler visits your site
  • Pages crawled: Which pages AI bots read most frequently
  • Crawl trends: Whether bot activity is increasing or decreasing over time

Why It Matters

If AI bots aren't crawling your site, they can't reference your latest content. If they're only crawling your homepage but ignoring your blog, your blog content won't influence AI responses. Bot tracking helps you understand the "supply side" of AI visibility.

Setup tip: You'll find the tracking script in the Site Traffic tab. It's a single script tag — takes 30 seconds to add to any website.

Suggestions or corrections

AI Focus Groups

3 min readTools & Features

AI Focus Groups let you simulate how different audience personas perceive your brand through AI. It's like running a traditional focus group — but instant, repeatable, and based on how AI actually represents you.

How It Works

Lighthouse generates synthetic personas based on your target audience — a "budget-conscious millennial," a "C-suite decision maker," an "early adopter" — and tests how AI responds to their specific types of questions about your industry and brand.

When to Use It

  • Before launching new messaging or positioning
  • When entering a new market segment
  • To test how different audiences might discover your brand through AI
  • To compare how AI describes you versus competitors from different audience perspectives
Strategy: Run focus groups before and after major content changes to measure the impact on AI perception. This creates a feedback loop for your content strategy.

Suggestions or corrections

Commerce Intelligence

3 min readTools & Features

Commerce Intelligence monitors how AI recommends products in shopping-related conversations. If you sell products, this shows whether AI assistants (including Amazon's Rufus) recommend you.

Key Metrics

  • Share of Shelf: What percentage of AI product recommendations include your brand
  • Average Rank: When your brand appears, where does it fall in the list?
  • Sentiment: How positively AI describes your products
  • Merchants: Which retailers AI associates with your products
For e-commerce brands: This is your equivalent of monitoring shelf space in retail stores — except the "shelf" is inside every AI conversation about your product category.

Suggestions or corrections

Reports & Exports

2 min readTools & Features

Lighthouse offers two types of export reports, accessible via the Export button in the dashboard header.

Full Brand Report

A comprehensive PDF covering your visibility across all AI platforms, sentiment analysis, citation mapping, competitive positioning, and strategic recommendations. Available to all users. Perfect for leadership presentations and stakeholder updates.

LLM Deep Dive Report (Admin only)

A detailed intelligence report focused on a single AI platform. Select the platform you want to analyze and Lighthouse generates a report with that platform's specific visibility, sentiment, citations, competitive landscape, and tailored recommendations. Ideal for platform-specific strategy development.

Pro tip: Both reports are designed for print-to-PDF. When the report opens, your browser will prompt you to save as PDF. The dark theme translates beautifully to print.

Suggestions or corrections

GEO vs. SEO: What's the Difference?

4 min readStrategy

If you've spent years optimizing for Google, you already have a strong foundation. But AI-driven discovery works differently than traditional search — and understanding those differences is key to your strategy.

The Core Difference

SEO optimizes for search engine results pages (SERPs) — the goal is to appear as a blue link when someone searches. GEO optimizes for generative engine responses — the goal is to be mentioned, cited, or recommended when someone asks an AI assistant.

What SEO and GEO Have in Common

  • Content quality matters in both
  • Authority and trustworthiness drive both rankings and AI mentions
  • Relevance to the user's query is essential
  • Technical foundations (fast site, proper structure) help both

Where GEO Differs

  • No rankings to check: There's no "position 1" in an AI conversation. Your brand either appears or it doesn't.
  • Context over keywords: AI understands meaning, not just keywords. Stuffing keywords won't help — having genuinely comprehensive content will.
  • Citations are the new backlinks: When AI cites your page, that's the GEO equivalent of a backlink.
  • Narrative matters: AI doesn't just list you — it describes you. The words around your brand name shape perception.
  • Multi-platform variability: You might be visible on ChatGPT but invisible on Claude. Each platform has different training data and behavior.
The bottom line: GEO isn't replacing SEO — it's an additional channel. The brands that invest in both will have the broadest discovery footprint.

Suggestions or corrections

How to Improve Your AI Visibility

5 min readStrategy

Improving your AI visibility isn't about gaming the system — it's about creating content that's genuinely useful, authoritative, and easy for AI to understand and reference.

1. Create Definitive Content

AI models love content that definitively answers questions. If you can become the "Wikipedia" for your niche — the most comprehensive, well-organized resource on your core topics — AI will reference you.

2. Structure for Citability

Use clear headings, bullet points, data tables, FAQ sections, and expert quotes. AI is more likely to cite content it can easily extract structured information from.

3. Build Third-Party Authority

Get mentioned in publications, industry reports, and directories that AI models trust. Check your citation sources in Lighthouse to see which publications AI already references in your space — then get featured there.

4. Keep Content Fresh

AI models are updated regularly, and they tend to favor recent, actively maintained content over outdated pages. Update your key pages at least quarterly.

5. Address the Prompts Where You're Missing

Look at your prompt-by-prompt analysis in Lighthouse. For every prompt where your brand wasn't mentioned, ask: "Do I have content that directly addresses this question?" If not, create it.

6. Monitor Competitor Behavior

If a competitor has higher visibility, study what they're doing differently. Are they publishing more frequently? Do they have content you don't? Are they cited by publications you're not mentioned in?

Timeline expectations: AI visibility improvements typically take 2-6 weeks to materialize, since models are updated on varying schedules. Be patient, be consistent, and use Lighthouse's scheduled analysis to track your progress.

Suggestions or corrections

Competitor Analysis

3 min readStrategy

Lighthouse automatically identifies which competitors AI mentions alongside your brand — and how often. Here's how to use that intelligence strategically.

Where to Find It

Competitor data appears in several places: the Overview tab shows a competitor frequency chart, the Prompts tab shows per-prompt competitor mentions, and the LLM Deep Dive reports include platform-specific competitive landscape breakdowns.

What to Do With It

  • Identify your real AI competitors: The brands that AI mentions alongside you may differ from who you consider your traditional competitors. Pay attention to who AI groups you with.
  • Find your competitive gaps: If a competitor appears in prompts where you don't, look at what content they have that you're missing.
  • Study citation sources: Which publications cite your competitors? Getting featured in those same sources can boost your visibility.
  • Track share-of-voice trends: Over time, is your competitor mention rate going up or down relative to yours?
Add competitors: When running an analysis, use the Advanced Options to add specific competitor URLs. Lighthouse will compare your visibility, sentiment, and citation data side by side.

Suggestions or corrections