All Posts
AEO

What an AI Visibility Audit Actually Covers for SMBs

Blog Image
Published on
June 5, 2026

You might rank on the first page of Google today. But when someone asks ChatGPT, Perplexity, or Google's AI Overviews to recommend a business like yours, there is a real chance your name never comes up. A 2026 study by SOCi analyzed more than 350,000 business locations across 2,751 brands and found that ChatGPT recommends just 1.2% of all local business locations. By comparison, those same brands appeared in Google's local 3-pack 35.9% of the time. That is a 30-times gap between traditional search visibility and AI search visibility, and it is the single most important number to understand before we go any further.

An AI visibility audit is the diagnostic process that maps where your brand currently stands in that gap. It tells you which AI platforms mention you, how accurately they represent you, which competitors are being recommended instead, and, critically, why. The findings are almost always fixable: the businesses that AI systems skip are usually missing specific structural signals, not deep authority, which means most SMBs are closer to a solution than they realize.

This guide covers exactly what a thorough audit examines, section by section. The research behind each point is cited and described so you can evaluate what you read on its merits, not on anyone's word.

Why your Google ranking and your AI visibility are two separate things

SOCi's research is blunt about the problem: in retail, only 45% of brands that lead in traditional local search also appeared among the most recommended in AI results. That means more than half of the businesses winning on Google are invisible when someone asks an AI assistant for a recommendation.

The reason is that AI search engines do not rank pages. They evaluate confidence. ChatGPT, Perplexity, and Gemini will exclude a source they are uncertain about rather than surface it with a caveat. That selectivity is why AI visibility is so concentrated, and why the businesses that get recommended tend to share a specific set of structural characteristics rather than just raw authority or backlink volume.

Consumer behavior is moving fast in the same direction. BrightLocal's 2026 Local Consumer Review Survey (1,002 US adult consumers) found that the share of consumers using AI tools like ChatGPT for local business recommendations climbed from 6% in 2025 to 45% in 2026. That puts AI tools in third place overall for local recommendations, behind only Google and Facebook. Meanwhile, 42% of those consumers now trust AI recommendations as much as traditional review platforms.

The gap, in plain numbers: Google's local 3-pack surfaces 35.9% of brands. ChatGPT surfaces 1.2%. Perplexity surfaces 7.4%. Gemini surfaces 11%. And 45% of consumers now use AI to find local services. An AI visibility audit is the tool that tells you which side of that gap you are on, and why.
Source: SOCi 2026 Local Visibility Index; BrightLocal 2026 Local Consumer Review Survey

‍

What an AI visibility audit actually covers

A serious audit is not a stack of ChatGPT screenshots. It is a structured diagnostic across at least five layers, each targeting a different reason AI systems might skip your business. Here is what each layer looks at and why it matters.

1. Baseline AI citation testing across multiple platforms

The first thing an audit does is run a structured set of prompts across the major AI platforms: ChatGPT (including the web-search-enabled version), Google AI Overviews and AI Mode, Perplexity, Microsoft Copilot, Gemini, and Claude. Each platform uses different source logic, so your visibility profile looks different on each one, sometimes dramatically so.

The prompt set is not five generic queries. A thorough audit tests 30 to 100 or more prompts organized by buyer intent stage: informational ("how does X work"), comparison ("X vs Y"), recommendation ("best X in [city]"), and transactional ("X near me accepting new clients"). The prompts are drawn from real customer language: your own search data, sales conversations, and industry forum research. This matters because the queries that expose AI visibility gaps are not always the obvious branded ones.

Research by AirOps and Kevin Indig, analyzing over 548,000 retrieved pages across 15,000 original prompts, found that ChatGPT generates an average of two or more follow-up "fanout" searches on 89.6% of queries, and that 95% of those fanout queries had zero monthly search volume by traditional keyword metrics. That means audits built only around the terms you already track can miss the majority of the queries where AI is actually making decisions about your brand.

What the baseline test produces: a platform-by-platform record of whether your business appears, whether the details are accurate, which competitors are being recommended in your place, and what sources the AI cited. That output becomes the benchmark for everything else in the audit.

One methodological note worth knowing: AI answers are probabilistic. The same prompt can produce different results across sessions. A credible audit runs each prompt multiple times and treats visibility as a rate, not a yes-or-no. Measurement should be repeated periodically, not treated as a one-time snapshot.

2. Entity and NAP consistency

"Entity" refers to how clearly and consistently AI systems can identify your business as a real, specific organization. AI platforms cross-reference your business information across Google Business Profile, Apple Maps, Bing Places, Yelp, Facebook, Foursquare, and major vertical directories. When your name, address, phone number, hours, category, or description conflicts across those sources, the model loses confidence in the listing and reduces how often it recommends you.

SOCi's 2026 data quantifies this directly: business profile accuracy on ChatGPT and Perplexity is only 68% for the brands they studied, compared to 100% accuracy on Gemini, which is grounded directly in Google Maps. The implication is clear: inconsistent data is one of the most common and most correctable causes of missed AI citations.

Whitespark's 2026 Local Search Ranking Factors report, based on a survey of 47 local search experts evaluating 187 ranking factors, added AI Search Visibility as a formal category for the first time. The panel found that citation signals account for 13% of the AI visibility category, and three of the top five factors for AI visibility relate to citations and consistent business information. Darren Shaw, Whitespark's founder, noted in the report:

"On-page signals are now the dominant factor for AI visibility at 24%, surpassing GBP signals. This is a wake-up call for businesses that relied solely on their Google profile."
Darren Shaw, Founder, Whitespark (2026 Local Search Ranking Factors Report)

‍

The entity check in an audit maps every major signal source and flags discrepancies. Even small inconsistencies, an abbreviated street name on one directory and a spelled-out version on another, can create enough ambiguity to depress citation frequency. Foursquare is one source that SMBs often overlook: its enterprise place data feeds directly into OpenAI's local layer, so a thin or inaccurate Foursquare listing can make a business invisible to ChatGPT regardless of how complete its Google profile is.

An audit also checks the quality and recency of your reviews across platforms. Review signals account for 16% of AI visibility weighting in the Whitespark framework, and the research team found that review recency, not just count, is what moves the needle: reviews older than 180 days carry only 10-20% of their original weight for AI recommendation purposes.

3. Schema markup and structured data

Schema markup is the structured layer that tells AI systems what your business does, who it serves, what questions it answers, and what qualifies it to answer them. An audit reviews whether LocalBusiness, Organization, FAQPage, HowTo, Review, and Service schema is present, correctly implemented, valid, and populated with accurate real-world data, not placeholders.

The evidence on schema's impact is real, though more nuanced than some vendors suggest. AirOps and Kevin Indig's study of 16,851 queries and 353,799 pages found that pages with JSON-LD schema had a 38.5% citation rate compared to 32.0% for pages without it, a 6.5 percentage point advantage. The AirOps 2026 State of AI Search Report found that about 61% of cited pages use three or more schema types, and pages with three or more types have a 13% higher likelihood of being cited than pages with one or fewer.

Schema type matters more than schema presence alone. FAQPage schema is the most consistently cited type across platforms, appearing in 10.5% of all cited pages in AirOps' study and achieving citation rates of 45.6% in Kevin Indig's controlled analysis. But the critical finding from AirOps is that FAQ schema without real Q&A content on the page has no effect. The model rewards the answer structure, not the metadata in isolation. Any page with FAQPage schema must have the question and a complete answer visible to the human reader.

The audit checks for what is present, what is valid, what is broken, and what is missing on the pages most likely to be cited, including service pages, location pages, and any page that directly answers a common customer question.

4. AI crawler accessibility

AI systems can only cite content they can actually read. This section of the audit checks whether the crawlers that power AI answer engines have clean access to your site, and it is one of the most commonly overlooked problems in AI visibility.

The major AI crawlers each have a specific user-agent string and respect your robots.txt file. The relevant crawlers for a standard audit are:

  • OAI-SearchBot and ChatGPT-User (OpenAI's search and retrieval agents for ChatGPT)
  • GPTBot (OpenAI's training and indexing crawler)
  • Claude-SearchBot, Claude-User, and ClaudeBot (Anthropic's retrieval and training agents)
  • PerplexityBot (Perplexity's indexing crawler)
  • Google-Extended (Google's crawler for Gemini model training)

A critical distinction: training crawlers (like GPTBot) and search/retrieval crawlers (like OAI-SearchBot and ChatGPT-User) are separate systems. Blocking GPTBot blocks training access to your content. Blocking OAI-SearchBot or ChatGPT-User means your content will not appear in ChatGPT search answers at all, even if GPTBot has already indexed it. OpenAI's own documentation states that sites opted out of OAI-SearchBot will not be shown in ChatGPT search answers. The two systems are independent.

Research from ziptie.dev found that approximately 27% of B2B and e-commerce sites unknowingly block major AI crawlers at the CDN layer (typically through Cloudflare's "Block AI Scrapers" toggle), even when their robots.txt looks correct. The CDN setting overrides robots.txt at the edge. Separately, several popular WordPress and Shopify SEO plugins added "block AI bots" toggles in 2024 and 2025 with the toggle enabled by default. A site owner who updated a plugin may have cut themselves off from ChatGPT, Claude, and Perplexity overnight without realizing it.

The technical crawlability review in an audit checks your robots.txt file for each of the crawlers listed above, checks your CDN and WAF configuration independently, checks your server logs for 403 responses to AI crawlers, and checks whether your site renders content server-side or relies on JavaScript execution that crawlers cannot complete. Client-side rendered sites often present only a shell to AI crawlers, leaving all the actual content invisible.

5. Content format and answer readiness

AI answers are built from short, extractable passages rather than entire articles. A page that buries the answer in a long preamble will be passed over in favor of a page that leads with a direct, quotable response. The audit evaluates whether your existing content is structured in a way that AI systems can extract and cite.

Kevin Indig's study found that pages with headlines that directly answer the question get cited 41% of the time, compared to 29% for pages with loosely related headlines. The same study found that the top 10-20% of a page generates the most citations across industries: ChatGPT cites heavily from the upper portion of a page, with the bottom 10% earning only 2.4-4.4% of citations. Conclusions are largely ignored.

AirOps' research also found that 83% of AI citations for commercial and evaluation-stage queries came from pages updated within the past 12 months, with more than 60% refreshed within the last six months. Stale pages fall out of citation rotation quickly. Once a fresher alternative is available, older content rarely regains visibility without a direct update.

The content review in an audit checks for direct-answer formatting (does the page answer a question in 40-60 words within the first two paragraphs), FAQ block presence and quality, heading structure (are H2 and H3 headings written as questions that match real customer language), named statistics with sources, and content freshness across the pages most likely to drive AI citations.

6. Third-party citation footprint

This is the component most SMBs underestimate, and the research makes clear it is one of the biggest score drivers. AI systems do not just read your website. They synthesize information from across the web, and the third-party sources they pull from for local and service businesses are specific and measurable.

BrightLocal's 2024 study analyzing 800 local searches across 20 verticals in 20 US cities found that ChatGPT Search pulls from business websites (58% of sources), business mentions in articles and press (27%), and directories (15%). Separately, research on citation patterns across AI platforms found that Wikipedia and Reddit together drive over 25% of ChatGPT citations in the US, and that YouTube correlation with AI visibility was the strongest single predictor in any 2025-2026 study at 0.737 (Ahrefs, 75,000-brand analysis).

For local service businesses, the audit maps your presence against the specific third-party platforms that AI engines actually pull from for your category: Google Business Profile, Yelp, Foursquare, Angi, Houzz, Healthgrades, Tripadvisor, Trustpilot, and industry-specific verticals. It also checks whether you have any presence on Wikipedia (if your business is notable enough for a stub), whether you appear in relevant Reddit threads or Quora discussions (organic participation, not promotion), and whether your YouTube channel, if you have one, is covering the questions your customers ask.

Sixty percent to 70% of local ChatGPT results pull from Foursquare's city guide listings, according to research by MediaELX. Foursquare is not a platform most SMBs actively manage, but it feeds directly into OpenAI's local data layer. A business with an incomplete or inaccurate Foursquare listing is effectively invisible to ChatGPT's local recommendation layer regardless of how strong its Google presence is. The citation map in the audit surfaces gaps like this.

‍

Which AI platforms get tested, and why testing only one is a mistake

The five platforms that matter most for SMBs right now are ChatGPT (including the web-search-enabled paid and free versions), Google AI Overviews and AI Mode, Perplexity, Microsoft Copilot, and Gemini. Each uses different sourcing logic, and your visibility profile can look very different across them.

Google AI Overviews are non-negotiable for SMBs because they appear directly on the search results page your current customers already use. BrightEdge's 12-month analysis tracking AI Overview presence from February 2025 to February 2026 found that AI Overviews now trigger on nearly half of all tracked queries, up 58% year-over-year. For specific industries, the numbers are higher: restaurant queries went from 10% to 78%, B2B Technology from 36% to 82%, and Education from 18% to 83%.

A key finding from BrightEdge's research on AI Overview citations: roughly 5 out of 6 AI Overview citations pull from content that is not on page one of traditional results. This varies by industry, from 24% overlap in Healthcare to 11% in Finance. Ranking well in organic search helps (pages at position one are cited 3.5 times more often than pages outside the top 20, per AirOps), but it does not guarantee AI Overview inclusion. The two systems operate on related but different logic.

Perplexity is worth prioritizing for B2B service businesses. Its sourcing behavior is more transparent than ChatGPT's, it shows citations explicitly in the interface, and it has strong adoption among professional buyers who prefer research-style answers with sources listed. Gemini tends to ground local answers directly in Google Maps, making it more responsive to Google Business Profile optimization. ChatGPT and Copilot draw from broader web indexes and are more influenced by third-party mentions, Wikipedia, and Reddit presence.

For businesses that depend on local walk-in traffic, voice assistants (Siri, Alexa, Google Assistant) should also be included in the audit scope. A customer asking their phone for "the best plumber near me" is using a different retrieval system than someone typing into ChatGPT, and the citation signals that power voice answers differ accordingly.

‍

‍How AI visibility gets scored and what the numbers mean

One of the most common complaints about AI visibility audits is that the scoring feels arbitrary. Understanding the metrics behind the score lets you evaluate any vendor proposal intelligently.

A credible audit uses at least five primary metrics to build a composite score:

  • Mention frequency: how often your brand appears across the full tested prompt set, expressed as a percentage.
  • Citation rate: how often your content or a page about you is used as a named source in the AI response, as distinct from a passing mention.
  • Share of voice: your mentions compared to a defined competitor set across the same prompt library. This number tells you whether you are the category leader, mid-pack, or trailing.
  • Accuracy score: whether the details AI surfaces about your business (name, location, hours, services, pricing) match your actual profile. SOCi's data showed 68% accuracy on ChatGPT and Perplexity for brands they studied. Knowing your number matters.
  • Position in response: whether your brand appears first in the AI answer or as a secondary mention. First-position citations receive significantly more user attention and, in platforms that show citations visually, a larger share of clicks.

‍

These signals are combined, often into a 0-100 index, and benchmarked against competitors in your category and market. A score in isolation means little without the benchmark: knowing you are at 22 is less useful than knowing the category leader is at 68 and the business you lose most deals to is at 41.

A strong audit also maps which intent stages you have coverage in and which you do not. Most SMBs show moderate presence on branded and informational queries and near-zero presence on recommendation-stage prompts like "what is the best [service] in [city]?" That stage is where the real decisions get made, and the gap between your informational visibility and your recommendation visibility is where remediation should start.

‍

‍What a quality audit delivers

Before hiring anyone to run an AI visibility audit, you should know what you will receive. A quality audit produces a specific set of documents. Any proposal that does not describe them clearly is a reason to ask harder questions.

The written report and visibility scorecard

The written report documents where your business appears across tested platforms and prompts, where it does not, what the AI systems say about you when they do mention you, and why specific competitors are being recommended in your place. The visibility scorecard translates the findings into a quantified, benchmarked assessment with enough context to understand whether your score is strong, average, or weak for your category and market.

The citation source map

This document lists the specific websites, directories, and publications that AI platforms pulled from when generating responses about businesses in your category, including screenshots, response transcripts, and source attribution by platform and query. It tells you not just what sources you need to appear in but also what third-party content may already be shaping your AI representation in ways you did not authorize or anticipate.

The competitor gap analysis

A competitor analysis earns its place in every quality deliverable set, not as a vanity exercise but as a prioritization tool. Knowing that a specific competitor appears in 71% of recommendation-stage prompts while you appear in 9% tells you exactly which intent stage to address first and what signals that competitor has that you do not. The deliverable should include which sources cite the competitor, what schema they use, what their review velocity looks like, and what content formats they publish that you do not.

The remediation roadmap

The roadmap translates all audit findings into ranked, sequenced implementation work: entity cleanup steps, schema fixes, content restructuring tasks, and citation gap targets. A quality audit tells you what to fix first and in what order based on the evidence from your own visibility data. That is the difference between a document you can act on immediately and one that sits in a folder.

‍

Scope, timeline, and realistic cost

Not every SMB needs the same depth of audit. Here is how to match scope to your actual situation.

A basic AI visibility audit

A basic audit is a one-to-three day engagement that tests two or three AI platforms with a focused prompt set and delivers a lightweight report covering your most obvious gaps and a starting priority list. Costs typically range from $250 to $1,000 depending on provider, market, and scope. This level works for a single-location SMB with a narrow service set and one primary market. It will not give you deep competitor benchmarking or a full citation source map, but it will surface the structural issues costing you the most citations right now.

An audit that covers only ChatGPT and ignores Google AI Overviews, Perplexity, and Gemini is not a complete AI visibility audit. Platform coverage should be listed explicitly in any proposal.

A comprehensive audit

A comprehensive audit is a one-to-two week project covering multi-platform benchmarking, full competitor analysis, a complete citation source map, a detailed content review, technical crawlability verification across platforms and CDN configuration, and a structured implementation roadmap delivered with a walkthrough session. Costs typically range from $500 to $1,500 or more for the audit itself, with implementation work separate. Rates vary by provider, market complexity, and scope.

This depth is appropriate for multi-location businesses, competitive markets, or regulated industries including healthcare, law, and finance where AI citation accuracy also carries compliance implications. It is also the right starting point for any business where buyers conduct significant research before making contact: if your sales cycle is longer than a week, the stages where AI is forming impressions matter more.

‍What to do with the findings: a 30-day action sequence

Audit findings are only useful if you act on them in the right order. Some fixes affect every AI platform simultaneously. Others take weeks to compound. Here is how to structure the first 30 days.

Week 1 and 2: Fix the structural signals first

Start with entity cleanup across directories and platform profiles. Audit your Google Business Profile, Apple Maps, Yelp, Foursquare, Facebook, and major industry directories for inconsistencies in your business name, address, phone number, category, and description. Prioritize the highest-impact discrepancies first and work through the list systematically. In parallel, implement or correct your schema markup, prioritizing LocalBusiness or Organization schema, then FAQPage on any page that directly answers a customer question, then Review markup where you have real aggregate rating data.

Check your robots.txt file and CDN configuration for unintentional blocks on OAI-SearchBot, ChatGPT-User, Claude-SearchBot, Claude-User, and PerplexityBot. If you are running Cloudflare or another CDN with an AI scraper blocking toggle, verify the setting in your dashboard independently of your robots.txt. These structural fixes carry the highest leverage because they affect every AI platform simultaneously. Initial movement on crawlability issues can appear in one to two weeks. Broader citation improvements typically emerge over 30 to 60 days.

Week 3 and 4: Restructure content and build citations

Identify the five to ten pages most likely to be cited for your highest-value service or product queries. For each page: rewrite the first paragraph to lead with a direct, 40-60 word answer to the most common customer question that page should address. Add an FAQ section using real customer language for the questions, with complete answers of two to four sentences each. Add named statistics with sources and dates where your expertise supports it. Update the last-modified date and refresh any figures or examples that are more than a year old.

Alongside the content work, begin targeted outreach to the specific citation sources your audit identified as gaps. Getting your business accurately listed or covered in the directories and publications that AI systems pull from for your category is the third-party signal that closes the loop between your owned content and your citation rate. The audit should tell you exactly which sources to target for your category and market.

A practical note on timeline: AI citation behavior changes continuously. BrightEdge's data showed that even among the most consistently cited domains, 87% of weekly changes were citation losses, not gains. This means an annual audit is too slow. Quarterly measurement is the minimum cadence for businesses where AI visibility materially affects customer acquisition.

‍

‍How to read your results: four diagnostic benchmarks

Once you have an audit report, these four patterns help you interpret the findings and prioritize correctly.

  • You appear in Gemini but not ChatGPT. Your Google Business Profile is strong, but your third-party citation footprint is weak. Foursquare, Wikipedia, Reddit, and YouTube presence are the likely gaps.
  • You appear in Perplexity but not ChatGPT. Your earned media and website content are working. The gap is likely your ChatGPT-specific signals: Wikipedia coverage, Reddit mentions, and YouTube presence.
  • You appear in ChatGPT but not Google AI Overviews. Your third-party citation footprint is solid, but your organic search ranking and on-page signals need attention. BrightEdge's 16-month research found that AI Overview citation overlap with organic top-10 grew from 32% to 54%, so organic ranking improvement directly lifts AI Overview inclusion over time.
  • Your mention rate stays below 5% after 60 days of entity and crawlability fixes. The bottleneck is off-site citations, not on-site structure. The solution requires building third-party visibility in the specific sources AI systems pull from for your category.

‍

The opportunity window

BrightEdge's April 2026 data found that only 19% of websites have specific directives for ChatGPT-related bots in their robots.txt files. The other 81% are treating AI search agents the same as traditional bots. Most SMBs have not begun this work, which means an audit performed now is closer to a first-mover diagnostic than a catch-up exercise.

For more context on how AI is affecting business strategy broadly, see Elementera's piece on The End of Easy AI: What the Latest Developments Mean for Your Business. For practical guidance on identifying which questions your buyers are actually asking AI systems, see the guide on how to do prompt research for AEO and AI search optimization. And for a deeper look at buyer intent, see user intent optimization and why it matters in AEO.

An AI visibility audit is a diagnostic baseline that makes every future content and technical decision more precise. Without it, you are optimizing blind, and the fixes you prioritize may not address the actual reason AI systems are skipping your brand.

The five components covered in this guide, entity consistency, schema markup, AI crawler access, answer-ready content, and third-party citations, form a complete picture of your AI search presence. Most businesses that go through this process find their gaps come from fixable structural problems rather than deep authority deficits. That is genuinely good news. You do not need to rebuild your digital presence from scratch. You need to make your existing presence machine-readable and citation-worthy.

Elementera runs structured AI visibility audits specifically for SMBs and delivers the complete deliverable set described in this guide, including a prioritized roadmap your team can implement without having to interpret raw data on your own. If you want to understand exactly where your business stands across AI search platforms, that is the right starting point.

‍

Sources

Sources cited in this article: SOCi 2026 Local Visibility Index; BrightLocal 2026 Local Consumer Review Survey; Whitespark 2026 Local Search Ranking Factors; BrightEdge AI Overviews at the One-Year Mark (2026); AirOps 2026 State of AI Search Report; Kevin Indig / AirOps ChatGPT Citation Study (16,851 queries, 353,799 pages); ziptie.dev B2B crawler blocking research; MediaELX Foursquare local data research; Contently AI Crawler Analysis (Q1 2026). All statistics cited from their primary or first-reference sources.