AI Search Citation Audit: How to Check If ChatGPT, Perplexity, and Gemini Are Quoting Your Brand

TL;DR

AI search engines like ChatGPT, Perplexity, Gemini, and Claude now influence buying decisions the same way Google results once did — but most brands have no idea whether they appear in AI answers at all. This guide gives you a concrete, repeatable audit protocol: 10 prompts tested across 4 AI engines, logged in a structured spreadsheet, scored into a single "citation share" metric. You can run the whole audit in one afternoon.

Key Takeaways

• AI citation visibility is measurable — but only if you test systematically across multiple engines and prompt types.
• Each AI engine has different citation behavior. Perplexity shows explicit source links. ChatGPT mentions brands by name but rarely links. Gemini pulls from a narrower index. Claude tends to be balanced but conservative.
• A 10-prompt × 4-engine audit (40 data points) is the minimum viable sample for a quarterly check.
• Your "citation share" — the percentage of relevant AI answers that mention your brand — is the single most important metric to track over time.
• Tools like BeRecommended can automate this process and track changes continuously.

Why You Need an AI Citation Audit

Traditional SEO gives you Google Search Console — impressions, clicks, average position. You know exactly where you stand. AI search gives you nothing. There is no "AI Search Console." No impressions dashboard. No click-through data.

This means most brands are flying blind in a channel that is growing fast. According to recent industry research, AI-assisted search now accounts for a meaningful share of product discovery queries, and that share is accelerating.

The problem is not that AI engines ignore your brand. The problem is that you do not know whether they do — and if they cite your competitor instead, you will never find out unless you actively test.

An AI citation audit fixes this. It gives you a baseline, a repeatable protocol, and a metric you can track quarterly.

The 4 AI Engines You Should Test

Each AI search engine behaves differently. Your audit needs to cover all four major players:

ChatGPT (with browsing enabled) — The largest user base. ChatGPT mentions brands frequently in conversational answers but rarely provides direct source links. Brand mentions here signal strong entity recognition.

Perplexity — The most transparent. Perplexity explicitly lists its sources with numbered citations. If your brand appears in Perplexity's citations, you have a verifiable, linkable data point. Reddit accounts for a significant portion of Perplexity's citations, so community presence matters here.

Google Gemini — Draws heavily from Google's own index and tends to favor established, high-authority sources. Gemini's citation behavior is narrower — it is harder to break into but more stable once you do.

Claude — Tends to be balanced and cautious with recommendations. Claude is particularly useful to test because it often gives nuanced, comparative answers rather than single-brand recommendations.

The 10-Prompt Audit Protocol

The key to a reliable audit is using consistent, categorized prompts. Here are 10 prompts across five intent categories, designed so you can adapt them to any industry by replacing the bracketed terms.

Category 1: Direct Brand Queries (2 prompts)

Prompt 1: "What is [Your Brand] and what does it do?"

Prompt 2: "Is [Your Brand] a good [product category] tool? What are the pros and cons?"

These test whether the AI engine recognizes your brand as an entity at all. If it returns accurate information, your entity footprint is solid. If it hallucinates or says it does not know you, that is your most urgent problem.

Category 2: Category Queries (2 prompts)

Prompt 3: "What are the best [product category] tools in 2026?"

Prompt 4: "Compare the top [product category] platforms for [target audience]."

These are the money prompts. When a potential customer asks for the best tool in your category, does your brand appear? In what position? These prompts have the highest commercial intent.

Category 3: Problem-Solution Queries (2 prompts)

Prompt 5: "How do I solve [problem your product solves]?"

Prompt 6: "What tools can help me with [specific use case]?"

These test whether AI engines associate your brand with the problem you solve — not just your category name, but the underlying pain point.

Category 4: Competitive Queries (2 prompts)

Prompt 7: "[Your Brand] vs [Competitor A]: which is better for [use case]?"

Prompt 8: "What are the best alternatives to [Competitor A]?"

These reveal how AI engines position you relative to competitors. You want to appear in "alternatives to" answers and win (or at least appear) in head-to-head comparisons.

Category 5: Thought Leadership Queries (2 prompts)

Prompt 9: "What are the latest trends in [your industry] in 2026?"

Prompt 10: "Who are the leading experts or companies in [your domain]?"

These test brand authority beyond product features. If AI engines cite you as a thought leader or trend-setter, your content strategy is working.

How to Run the Audit: Step by Step

Step 1: Set up your tracking spreadsheet. Create a spreadsheet with columns for: Prompt Number, Prompt Text, Engine, Brand Mentioned (Yes/No), Position (1st, 2nd, 3rd, or Not Listed), URL Cited (if any), Competitor Brands Mentioned, Timestamp, and Notes.

Step 2: Run each prompt across all 4 engines. Open ChatGPT, Perplexity, Gemini, and Claude in separate browser tabs. Enter the same prompt in each. Log the results immediately — AI answers can vary between sessions, so capture them in real time.

Step 3: Use fresh sessions. Start a new conversation for each prompt. Previous context in the same chat can bias the results. For ChatGPT and Claude, explicitly start a new conversation. For Perplexity, use the fresh search bar.

Step 4: Run it twice. AI responses are non-deterministic. Run the full 40-test matrix on two separate days (ideally 3-7 days apart) and average the results. This smooths out random variation.

Step 5: Score your citation share. Count the total number of tests where your brand was mentioned. Divide by the total tests (40 per run, 80 if you ran twice). That percentage is your citation share.

Example: Your brand appeared in 12 out of 80 tests = 15% citation share.

How to Interpret Your Results

0-5% citation share — Your brand is effectively invisible to AI search. Priority: build entity recognition through Wikipedia, industry directories, review platforms (G2, Capterra, Trustpilot), and authoritative third-party mentions.

5-15% citation share — You are on the radar but inconsistent. Look at which prompt categories you appear in and which you miss. Typically, brands at this level appear in direct brand queries but are absent from category and competitive queries.

15-30% citation share — Solid emerging presence. You are likely appearing in some category queries. Focus on increasing consistency and expanding into problem-solution and thought leadership categories.

30%+ citation share — Strong AI visibility. You are a recognized entity in your category. Focus shifts to maintaining position, monitoring competitors, and optimizing for new prompt patterns.

Automate It: From Manual Audit to Continuous Monitoring

The manual audit gives you a baseline. But AI answers change constantly — new training data, updated retrieval indexes, shifting citation patterns. A quarterly manual audit catches big shifts but misses the daily dynamics.

BeRecommended automates this entire workflow. It runs your prompt set across AI engines continuously, tracks citation share over time, alerts you when competitors gain or lose visibility, and gives you a dashboard instead of a spreadsheet. If you are serious about AI search visibility, start with the manual audit to understand the mechanics, then move to automated monitoring to stay ahead.

Common Mistakes to Avoid

Testing only one engine. ChatGPT visibility does not mean Perplexity visibility. Each engine has different data sources and citation behavior. Always test all four.

Using leading prompts. Do not ask "Why is [Your Brand] the best?" — that biases the answer. Use neutral, customer-like language.

Ignoring the competitors column. Your audit is not just about whether you appear. It is about who appears instead of you. Track competitor mentions with the same rigor.

Running the audit once and forgetting. AI citation landscapes shift faster than Google rankings. Commit to quarterly audits at minimum.

FAQ

Q: How long does the full manual audit take? A: About 2-3 hours for the initial 40-test run, including setup and logging. The second verification run takes about 90 minutes since your spreadsheet is already set up.

Q: Do I need API access to run this audit? A: No. The manual protocol uses the free web interfaces of all four AI engines. API access is only needed if you want to automate at scale.

Q: How often should I repeat the audit? A: Quarterly for the full 10-prompt protocol. Monthly if you are actively running AI visibility campaigns and need faster feedback loops.

Q: Does this work for local businesses, not just SaaS? A: Yes. Adapt the prompts to include location context: "best [service] in [city]" or "[service] near [location]." AI engines increasingly handle local intent.

Q: What if my brand does not appear in any AI answer? A: Start with entity building. Get listed on Wikipedia (if notable enough), major review platforms, and industry directories. Publish content that directly answers the questions in your audit prompts. Read our guide to Generative Engine Optimization for the full framework.

Q: Can I use BeRecommended to automate this? A: Yes. BeRecommended runs citation audits continuously across AI engines, tracks your citation share over time, and benchmarks you against competitors. It replaces the spreadsheet with a live dashboard.

Sources

• Profound AI citation analysis (27M citations across ChatGPT, Gemini, AI Overviews)
• 5W AI Platform Citation Source Index 2026
• Otterly.ai AI search monitoring methodology
• BeRecommended internal benchmark data, Q1-Q2 2026
• What Is Generative Engine Optimization (GEO)?
• AI Citation Source Index 2026: What Brands Should Do
• AI Brand Monitoring and Reputation Guide
• Prompt-Level SEO Experiments: Testing Framework for GEO