How to Run an AI Visibility Audit: A Step-by-Step Guide

Ask ChatGPT to recommend a vendor in your industry. If your brand doesn't appear, you have an AI visibility problem — and you're not alone. According to Onely's GEO research, keywords triggering AI Overviews grew from 1.5% to roughly 32% within 12 months — a 20× expansion of AI-mediated search. Meanwhile, visitors arriving through AI search convert at 4.4× the rate of traditional organic traffic.

The question is no longer whether AI search matters. It's whether your brand is visible in it. An AI visibility audit gives you the answer: a systematic assessment of how your brand appears across ChatGPT, Perplexity, Gemini, and other AI search engines, revealing exactly where you stand and where to improve.

What Is an AI Visibility Audit?

An AI visibility audit is a structured process of testing branded and category prompts across major AI search engines to measure how often, how prominently, and how accurately your brand appears in AI-generated answers. It is the diagnostic foundation of any Generative Engine Optimization (GEO) strategy — without it, optimization is guesswork.

Unlike a traditional SEO audit, which examines keyword rankings and indexation status, an AI visibility audit measures something fundamentally different: whether AI engines cite you when users ask questions. This distinction mirrors the broader difference between GEO and SEO — traditional search is about ranking; AI search is about being referenced.

The 6 Core Metrics of an AI Visibility Audit

Based on industry frameworks (including Visiblie's AI visibility metrics research), a comprehensive audit tracks six dimensions:

An important caveat: there is no universal "good" AI visibility score. The benchmark depends on your category, competitor set, and the prompts you test. Treat scores as trends to track over time, not absolute grades.

10 Steps to a Complete AI Visibility Audit

Step 1: Define Your Audit Scope

Decide which brands, product lines, and markets to cover. B2B companies operating in both domestic and international markets need to test across language-specific AI engines — DeepSeek and Baidu's Wenxin for Chinese markets, plus ChatGPT, Perplexity, and Gemini for global audiences.

Step 2: Build Your Prompt Library

The quality of your audit depends entirely on the quality of your prompts. Create 30-50 test prompts per product line across four categories: brand queries ("What is [brand]?"), category searches ("best [product type] for [use case]"), scenario-based questions ("how to solve [problem]"), and competitor comparisons ("[brand A] vs [brand B]").

Step 3: Collect Baseline Data

Run each prompt through your target AI platforms and record: whether your brand was mentioned, whether it was recommended, its position in the response (first, middle, or last), factual accuracy of the mention, and which competitors appeared alongside you. Manual testing works for small audits; automated tools like AiVisible are necessary at scale.

Step 4: Verify Crawler Access

Check that your robots.txt doesn't block AI crawlers. Key bots to verify include GPTBot (OpenAI), ClaudeBot (Anthropic), PerplexityBot, and Google-Extended. If AI search engines can't crawl your content, no amount of content optimization will help.

Step 5: Assess Content Citability

AI engines favor content that is clearly structured, information-dense, and easy to extract. Research shows that content clarity and summarizability correlate with a 32.83% increase in citation likelihood, while strong E-E-A-T signals contribute a 30.64% lift (per Onely's analysis). Critically, 55% of AI citations come from the first 30% of page content — front-load your key information.

Step 6: Audit Structured Data

Review your Schema markup implementation. While 71% of websites deploy some form of structured data, only 22% pass Google's Rich Results Test cleanly. Focus on Organization, Article, FAQPage, and Product schemas — correct implementation correlates with a 21.60% improvement in citation probability.

Step 7: Scan Third-Party Sources

AI engines don't just reference your website. Wikipedia, YouTube, Reddit, industry forums, and review platforms all feed into AI knowledge bases. Third-party platforms account for approximately 38% of AI Overview citations. Verify that brand information across these channels is accurate and consistent with your official messaging.

Step 8: Benchmark Against Competitors

Run the same prompts for 3-5 key competitors and calculate relative share of voice. Pay attention to which platforms and scenarios give competitors an edge — these gaps often reveal your highest-impact optimization opportunities.

Step 9: Generate the Audit Report

Compile your findings into an actionable report covering: per-platform metric scores, competitor benchmarking, key strengths and weaknesses, and prioritized recommendations. The report should answer one question clearly: "Where should we invest to maximize AI visibility improvement?"

Step 10: Build Your Optimization Roadmap

Rank optimization actions by impact and implementation difficulty into three tiers: Critical (fix immediately), High Priority (this month), and Medium/Lower (this quarter). For practical optimization tactics, see our guide on AI search marketing for manufacturers.

What Drives AI Citation: 6 Factors That Matter

When AI search engines decide whether to cite your content, six factors carry the most weight. Based on analysis of 36 million AI Overviews:

Source: Onely GEO Checklist Research, 2026

Three Common Audit Mistakes

Mistake 1: Drawing conclusions from a single test. AI search results are inherently probabilistic — the same prompt can produce different answers at different times. One data point is a snapshot, not a trend. Run at least three rounds of testing before making strategic decisions.

Mistake 2: Blending platform data into a single score. Each AI engine uses different RAG pipelines and knowledge sources. Your brand may perform well on ChatGPT but be absent from Perplexity entirely. Aggregated scores hide the real problems — always track per-platform metrics.

Mistake 3: Auditing in isolation without competitor context. AI visibility is relative. When an AI engine recommends your brand, it's simultaneously choosing not to recommend others. Without competitive benchmarking, you can't tell whether your visibility is genuinely strong or merely present.

Frequently Asked Questions

How long does an AI visibility audit take?

A thorough audit typically takes 1-3 days. Data collection across platforms requires a few hours for manual testing (or minutes with automated tools like AiVisible), while analysis and reporting need 1-2 days. Start with a focused audit of 20-30 prompts and expand from there.

How many prompts should I test?

Industry best practice suggests 30-50 prompts per core product line, covering brand queries, category searches, use-case scenarios, and competitor comparisons. For an initial audit, 20 well-crafted prompts provide a meaningful baseline.

How often should I run an AI visibility audit?

Monthly at minimum. AI search results are probabilistic, so a single check is a snapshot. Only consistent monthly monitoring reveals reliable trends and the impact of your optimization efforts.

Does my brand show up differently across AI platforms?

Almost always, yes. Different AI search engines use different RAG pipelines and knowledge sources. A brand that appears prominently in ChatGPT responses may be completely absent from Perplexity or Gemini. Platform-specific tracking is essential.

How is this different from a traditional SEO audit?

Traditional SEO audits focus on keyword rankings, indexation, and technical page metrics. AI visibility audits measure whether AI engines mention, recommend, and cite your brand. The two are complementary: SEO ensures search engines can crawl your content, while GEO ensures AI engines choose to reference it. Read more in our GEO vs SEO comparison.