Your rankings dashboard cannot tell you whether ChatGPT, Gemini, Perplexity, or Google’s AI Overviews ever mention your brand, and for a growing share of buyers those answers are the only shortlist that exists. This guide walks through a free SEO audit you can run yourself in an afternoon: interrogate the engines with a structured prompt panel and score every response, confirm AI crawlers can reach and render your site, then grade whether your pages contain anything an answer engine can safely lift and cite. You will end with a real AI search visibility baseline, a list of specific gaps, and a way to measure progress monthly instead of guessing.
The managing partner of a regional accounting firm lost a five-figure engagement this spring and only learned why during the courtesy call afterward. The prospect had asked ChatGPT to shortlist firms with the right industry experience, gotten four names back, and interviewed those four. The firm that lost was not evaluated and rejected. It was never in the room. Its Google rankings had not moved an inch the entire time.
That is what AI search invisibility looks like from the inside: nothing. No alert fires, no ranking drops, no error appears in any report. The only way to find out is to go looking deliberately, and that is exactly what this guide is for. What follows is a free SEO audit built for the AI era, three stages you can run without buying a single tool, plus an honest accounting of where the do-it-yourself version stops and what sits beyond it.
How Big Is the Blind Spot You Cannot See?
The scale is no longer debatable. Google reported in mid 2025 that AI Overviews reach more than 2 billion users a month, with its dedicated AI Mode passing 100 million monthly users in the US and India shortly after launch. ChatGPT and Perplexity add their own enormous query volume on top, none of it visible in the ranking trackers most businesses still treat as the scoreboard.
The behavior shift matters as much as the reach. When an AI summary answers the question directly, fewer people ever touch the traditional results underneath it, which means a brand can hold its blue-link positions while the audience quietly stops arriving through them.
Pew Research Center found that Google users clicked a traditional result on just 8 percent of visits where an AI summary appeared, compared with 15 percent when it did not. The click you have spent years optimizing for happens roughly half as often once an AI answer stands in front of it.
There is a counterweight, and it is the reason this audit is worth an afternoon. Semrush’s research on AI search traffic suggests visitors arriving from AI experiences are roughly 4.4 times more likely to convert than visitors from traditional search, because the answer engine has already done the qualifying before the click happens. Fewer visits, dramatically better visits. The brands collecting them are the ones the engines can see, which is why AI search visibility deserves the same measurement discipline that rankings have enjoyed for twenty years.
What Are the Warning Signs of an Invisible Brand?
Invisibility leaves fingerprints, they are just scattered across places nobody assigns anyone to watch. Prospects arrive later in their decision than they used to, with a shortlist already formed somewhere you cannot audit. Sales calls include phrases like “we asked ChatGPT” or “Perplexity suggested.” Lead volume drifts down over two or three quarters while rankings and traffic reports insist everything is stable. A competitor with objectively weaker organic positions keeps showing up in deals anyway. Each of these is an AI search visibility problem wearing a different costume, and none of them appears in a rankings report.
Misdescription is the subtler sibling of absence, and in our client work it turns up almost as often. The engine does mention the brand, but describes a service line that was discontinued, an old market focus, or pricing positioning that is years out of date. A wrong answer delivered confidently can cost more than no answer, and it is one of the specific failure modes that shaped how we think about AI-era search in the first place.
Stage One: What Do the Engines Say When You Ask About You?
Build a prompt panel before you type anything, because ad hoc testing produces ad hoc conclusions. A working panel needs three layers. Branded prompts establish the accuracy baseline: what is the company, what does it do, who does it serve. Category prompts are where deals live: best providers of your service, in your market, for your buyer type, phrased the way a real customer would phrase it. Comparison prompts test the shortlist moment directly: your brand against the two competitors you lose to most, and open-ended “who should I choose for X” questions where you are not named at all.
Run the panel across ChatGPT, Gemini, Perplexity, and Google with AI Overviews, logged out where possible and without leaning on chat history, since personalization will otherwise flatter you. Geography changes the panel more than most teams expect. A single-market firm should phrase prompts the way a local buyer would, while brands competing everywhere need prompts sampled across regions, which is the same coverage logic that shapes national SEO campaigns on the traditional side of search.
Score every response into one of four buckets and record it: cited with a link, mentioned without a link, absent, or misdescribed. A spreadsheet with the prompt, the platform, the date, the bucket, and which competitors appeared is the entire tooling requirement. By the end you will have the number most businesses have never seen: the percentage of high-intent questions in their category that they appear in at all. Note the competitors named in every response too, since the same spreadsheet doubles as a share-of-voice record, and watching a rival’s name climb across your own panel is usually the argument that finally gets AI search visibility onto the leadership agenda.
A grid concept with the four scoring outcomes as columns, cited, mentioned, absent, misdescribed, and the four platforms as rows, with each cell shaded by how often that outcome occurred during the audit. The visual point is that the pattern diagnoses the problem: a row that is mostly absent points to a reachability or authority gap on that one platform, while a column of misdescribed cells across every platform points to stale or conflicting information about the brand itself.
Stage Two: Can the Machines Reach and Render Your Site?
Open yourdomain.com/robots.txt and read it line by line against the AI user agents that matter: OAI-SearchBot for ChatGPT search visibility, PerplexityBot, Google-Extended, and the various Bing agents. Plenty of sites blocked everything AI-related during the scraping backlash of 2023 and 2024, sometimes through a plugin update nobody reviewed, and a block written for training bots that also catches search bots is functionally a request to be invisible.
Rendering is the check almost every do-it-yourself audit skips. Most AI crawlers read the raw HTML response and execute little or no JavaScript, so a site that builds its content client-side can look complete in a browser and nearly empty to the machine deciding whether to cite it. View the page source directly, or fetch the page with a command line tool, and confirm your actual copy appears in the initial HTML. Overaggressive bot protection deserves the same scrutiny, since firewall rules and challenge pages tuned for scrapers routinely swallow legitimate AI crawlers too. These infrastructure checks are the unglamorous core of the diagnostic-led SEO services we run before any content recommendation gets made, because content advice is worthless on a site the engines cannot read.
Stage Three: Is There Anything on Your Pages Worth Quoting?
Answer engines are quotation machines. They lift the clearest, most self-contained, most attributable statement available and build around it, which means the audit question for every important page is brutally simple: if a machine could quote exactly one passage from this page, does a passage worth quoting exist? Pull up your five most valuable pages and read only the first two sentences under each heading. If those sentences set context instead of stating answers, the page is structured for a reader who no longer arrives.
The same pass should grade freshness and identity. Pages with no visible dates, statistics with no stated source, and an about page that says less about the company than its LinkedIn profile does are all quiet disqualifiers. The scorecard below condenses the full three-stage audit into the checks that separate visible brands from invisible ones. This whole stage is also the most realistic starting point for smaller teams, and it is where the SEO programs we build for small businesses tend to find the fastest wins, since content fixes need no engineering queue.
| Audit Area | Passing Looks Like | Failing Looks Like |
|---|---|---|
| Prompt panel results | Cited or accurately mentioned in most category prompts | Absent from open-ended prompts, or described inaccurately |
| Crawler access | Search-focused AI bots explicitly allowed in robots.txt | Blanket AI blocks, or firewall challenges stopping crawlers |
| Rendering | Core copy present in the raw HTML response | Content assembled by JavaScript after page load |
| Quotability | Direct, self-contained answers in the first sentences under headings | Long preamble before any extractable claim |
| Trust signals | Dated pages, sourced statistics, substantive about page | Undated content, unsourced claims, thin company identity |
Where Does the Do-It-Yourself Audit Hit Its Ceiling?
Everything above is real and worth doing, and it has honest limits. A hand-run prompt panel samples a couple dozen questions out of the thousands of phrasings real buyers use, so a clean result can coexist with large invisible gaps. Server log analysis, which is the only way to confirm what AI crawlers are doing on your site rather than what robots.txt permits them to do, requires access and fluency most marketing teams do not have on hand. And the free audit tells you where you stand, not why: separating a reachability problem from an authority problem from an entity confusion problem is diagnostic work, and treating the wrong one burns quarters.
A free audit is worth running precisely because it is free and fast. Its job is not to fix your AI search visibility. Its job is to end the era of not knowing, which is the most expensive era a brand can be in.
The following math is illustrative, with every assumption stated, so treat the structure as the takeaway rather than the specific figures. Assume a competitor’s visibility earns them 400 AI-referred visits a month in your category. Assume your organic visitors convert at 2 percent, and apply Semrush’s finding that AI search visitors are roughly 4.4 times more likely to convert, which puts those visits near 8.8 percent. That is around 35 conversions a month, and at a modest 2,500 dollars average value, roughly 88,000 dollars in monthly opportunity flowing to a brand the engines can see. Your real numbers will differ. The asymmetry between highly qualified AI traffic and its invisibility to your current reporting will not.
Across the audits we have run since AI referrals became measurable, the gap between what a team assumed and what the prompt panel showed has been the single most consistent finding, and closing it professionally is not the budget shock many teams brace for. Our SEO pricing and plans are published openly for exactly that reason, since the honest comparison is never audit versus no audit, it is the cost of visibility work against the cost of staying absent from the answers buyers already trust.
How Do You Keep Score After the First Audit?
A one-time audit decays fast, because answer engines refresh their picture of brands continuously. Rerun the same prompt panel monthly, on the same platforms with the same scoring, and track one headline number: the share of category prompts where you are cited or accurately mentioned. That is your AI search visibility score, and its trend line matters far more than any single month’s reading. Annotate the timeline with what shipped and when, content rewrites, schema updates, robots.txt changes, so movements can be traced back to causes instead of argued about.
Underneath it, watch three supporting layers. Referral sessions from chatgpt.com, perplexity.ai, gemini.google.com, and copilot.microsoft.com, segmented separately in analytics, tell you when mentions start converting into measurable visits. Competitor share within your own panel tells you whether you are gaining ground or merely holding it. And the misdescription count deserves its own line, since accuracy regressions after a rebrand, a service change, or a site migration are common and correctable if caught early. Teams comparing how agencies handle this kind of ongoing accountability can see why brands choose Skyfield for the reporting side specifically, because a visibility program without a scoreboard is indistinguishable from hope.
Frequently Asked Questions
How do I check if my brand shows up in AI search results?
Run a structured prompt panel: branded, category, and comparison questions asked across ChatGPT, Gemini, Perplexity, and Google’s AI Overviews, logged out to limit personalization. Score each response as cited, mentioned, absent, or misdescribed, and record the results. The share of category prompts where you appear accurately is your baseline AI visibility score.
Why does my site rank well on Google but never get mentioned by ChatGPT?
The systems are separate. ChatGPT relies on its own crawling and retrieval rather than Google’s index, so a robots.txt block, a JavaScript-heavy site, weak brand corroboration across the wider web, or content with nothing quotable can all keep a well-ranked site out of its answers. The audit’s second and third stages are designed to isolate which cause applies.
Is AI search traffic actually worth pursuing yet?
The volume is smaller than traditional organic for most sites, but the quality runs far higher. Semrush’s research suggests AI search visitors are roughly 4.4 times more likely to convert, since the engine pre-qualifies them before the click. Treating it as a small channel because the session count is small misreads what the sessions are worth.
What does a free SEO audit for AI visibility actually include?
The do-it-yourself version covers three stages: a scored prompt panel across the major engines, a technical pass confirming AI crawlers can access and read your site, and a content review grading whether your key pages contain direct, quotable, dated answers. A professional audit adds server log analysis, entity checks, and competitive share-of-voice measurement at scale.
How often should I re-run an AI visibility audit?
Monthly is the practical cadence for the prompt panel, since answer engines update their view of brands continuously and single readings are noisy. The technical access checks only need revisiting after site changes, migrations, plugin updates, or firewall adjustments, which are the moments crawler blocks get introduced by accident.
Can a small business compete with big brands in AI answers?
Often more effectively than in traditional rankings. Answer engines reward specificity, and a small firm with a tight specialty, consistent information across the web, and genuinely quotable content can win prompts phrased around its exact niche even when broad head terms remain out of reach. The prompt panel usually reveals which niche questions are winnable first.
What should I fix first if the audit shows I am invisible?
Sequence by dependency. Access problems come first, since nothing else matters while crawlers are blocked or the content is invisible to them. Accuracy problems come second, because a misdescribed brand is actively losing trust. Quotability and content depth come third, and they are the layer that compounds once the first two are clean.
We will run the full professional version of this audit on your brand, prompt panel, crawler access, content grading and all, at no cost.
Sources
| Pew Research Center | Google Users Are Less Likely to Click on Links When an AI Summary Appears in the Results |
| TechCrunch | Google’s AI Overviews Have 2B Monthly Users, AI Mode 100M in the US and India |
| Semrush | Semrush AI Search Study: What 10M Keywords Reveal About Traffic and Conversions |
| Google Search Central | AI Features and Your Website |
| Search Engine Land | Google’s AI Overviews Are Hurting Clicks: Pew Study |
