Skip to content

AI Visibility Audit

Fresh

This content reflects a synthesized audit workflow from Metehan Alp's methodologies as of 2025-2026.

A systematic audit workflow to assess your current visibility across AI search platforms. Run this before building an optimization strategy -- you need a baseline before you can measure improvement.

Audit Overview

Step 1: CC Rank Check

Establish your domain's citation authority baseline using Metehan's CC Rank Checker tool.

Process

  1. Open the CC Rank Checker at metehan.ai
  2. Enter your domain
  3. Record the citation rank score
  4. Run the check for your top 3 competitors as well
  5. Document the gap between your score and the top competitor

What You Learn

  • Your domain's relative citation authority
  • How you compare to competitors in AI citation strength
  • Whether your domain is even in the running for AI citations

Output

  • Domain CC rank score
  • Competitor CC rank scores (top 3)
  • Gap analysis (numeric difference)

Step 2: Temperature Zero Brand Audit

Test what AI platforms currently say about your brand when asked directly.

Process

  1. Open ChatGPT, Perplexity, Claude, and Google AI Mode
  2. For each platform, ask these 5 prompts:
    • "What is [Brand Name]?"
    • "Is [Brand Name] good at [your primary service]?"
    • "Compare [Brand Name] to [Top Competitor]"
    • "What do people say about [Brand Name]?"
    • "Who is the best [your category] company?"
  3. Record each response verbatim
  4. Score each response: Positive / Neutral / Negative / Not Mentioned

What You Learn

  • Whether AI platforms know your brand exists
  • What they say about you (accuracy, sentiment, completeness)
  • Whether competitors are mentioned instead of you
  • Specific inaccuracies that need correction

Output

  • 20 recorded responses (5 prompts x 4 platforms)
  • Sentiment scoring grid
  • Inaccuracy list
  • Competitor mention tracking

Step 3: Log Probability Analysis

Analyze how confidently AI models associate your brand with your topic area.

Process

  1. Use an AI model that exposes token probabilities (OpenAI API with logprobs enabled)
  2. Submit prompts like: "The leading company for [your service] is"
  3. Record the log probability assigned to your brand name as the next token
  4. Compare against competitors' brand names for the same prompt
  5. Test across 10-15 prompt variations

What You Learn

  • How strongly the AI model associates your brand with your category
  • Whether competitors have stronger token-level associations
  • Which topic areas your brand is most/least associated with

Output

  • Log probability scores per prompt per brand
  • Ranking of brands by association strength
  • Weak association areas (optimization targets)

Step 4: Server Log Review

Analyze your server logs for AI bot crawl patterns.

Process

  1. Pull server logs for the last 90 days
  2. Filter for known AI bot user agents:
    • GPTBot
    • OAI-SearchBot
    • GoogleOther
    • PerplexityBot
    • CCBot
    • ClaudeBot
  3. For each bot, record:
    • Total requests
    • Pages crawled (unique URLs)
    • Crawl frequency (requests per day)
    • Most-crawled pages (top 20)
    • wp-json and sitemap access patterns
  4. Verify bot identity using reverse DNS (see Bot Crawlers Reference)

What You Learn

  • Which AI bots are crawling your site
  • How frequently they crawl
  • Which pages they prioritize
  • Whether wp-json endpoints are being accessed (1.6x GPTBot activity signal)
  • Any blocked bots that should be allowed

Output

  • Bot traffic summary table
  • Top 20 most-crawled pages per bot
  • Crawl frequency trends (daily/weekly)
  • Blocked bot list review

Step 5: Prompt Tracking

Transform your GSC data into AI prompt intelligence using the AI Prompt Tracking methodology.

Process

  1. Export GSC query data (last 90 days, all queries with impressions > 10)
  2. Filter for conversational queries (4+ words, question format, comparison structure)
  3. Cluster by semantic similarity using embeddings
  4. Generate AI prompt variants from clusters
  5. Validate top 50 prompts against ChatGPT, Perplexity, and Google AI Mode
  6. Record citation status for each validated prompt

What You Learn

  • Which prompts you should be tracking (data-driven, not guessed)
  • Current citation rate across platforms
  • Highest-value prompt targets (high GSC impressions + not yet cited)

Output

  • Prompt cluster list (23+ clusters typical)
  • Citation status grid (prompt x platform)
  • Priority targets (high impression, non-cited prompts)

Step 6: Platform Coverage

Synthesize findings into a platform-by-platform coverage assessment.

Process

For each platform (ChatGPT, Perplexity, Google AI Mode), compile:

  1. Brand awareness: Does the platform know your brand? (from Step 2)
  2. Citation frequency: How often are you cited? (from Step 5)
  3. Bot crawl activity: Is the platform's bot crawling you? (from Step 4)
  4. Association strength: How strongly does the platform link your brand to your topic? (from Step 3)
  5. Content coverage: What percentage of relevant prompts cite your content? (from Step 5)

Coverage Scoring

MetricScore
Brand recognized + cited in 50%+ of promptsStrong
Brand recognized + cited in 20-49% of promptsModerate
Brand recognized + cited in < 20% of promptsWeak
Brand not recognizedNot Present

Output

  • Platform coverage scorecard
  • Platform-specific optimization priorities
  • Cross-platform gap analysis

Audit Report Template

Compile all outputs into a single audit report:

AI VISIBILITY AUDIT: [Brand Name]
Date: [Date]

1. CC RANK: [Score] (Competitor avg: [Score])
2. BRAND AWARENESS: [Platform coverage summary]
3. CITATION RATE: [X]% across [N] validated prompts
4. BOT ACTIVITY: [Summary of crawl patterns]
5. PRIORITY TARGETS: [Top 10 optimization opportunities]
6. RECOMMENDED NEXT STEP: [Workflow recommendation]

Next Steps After the Audit

Audit FindingRecommended Workflow
Low citation rate, good content existsAEO Content Pipeline (optimize existing)
Content gaps in key topic areasTopic Cluster Building
Bots not crawlingFix technical access (robots.txt, sitemap, wp-json)
Brand not recognizedBuild brand signals (schema, E-E-A-T, CiteMET distribution)