What Is a GEO Audit, and How Do You Run One in 2026?

68% of US Google searches now end without a click (SparkToro + Similarweb, early 2026), up from 60% just two years earlier, and a fast-growing share of those zero-click answers are written by ChatGPT, Perplexity, Gemini, and Google's AI Overviews. If those engines are quoting your competitor instead of you, you are losing pipeline you cannot even see in Google Analytics. A GEO audit is how you make that invisible loss visible.

A GEO (Generative Engine Optimization) audit is a forensic check of whether AI answer engines cite your website for your buyers' questions, which competitors they cite instead, and what is missing from your site (llms.txt, schema, answer-ready content) that would earn those citations back.

Here is the failure most teams hit first: they check their Google rankings, see them holding steady, and conclude nothing is wrong. Meanwhile their AI-citation rate is near zero and they have no dashboard that would ever tell them. This post walks through what a GEO audit measures and the exact 7 steps to run one.

What does a GEO audit actually check?#

A GEO audit checks four things: whether AI engines cite your domain for your real buyer queries, which competitors get cited in your place, whether AI crawlers can even read your site, and whether your content is structured as answers an engine can lift. It is a citation test plus an infrastructure test, not a keyword-ranking report.

Traditional SEO asks "where do I rank on the Google results page?" A GEO audit asks a different question: "when a buyer asks an AI assistant about my category, does the assistant name me?" Those are not the same thing. You can rank page-one on Google and be invisible inside ChatGPT, because the engines build answers from sources they trust and can parse, not from the top blue links.

Why is a GEO audit different from an SEO audit?#

A GEO audit measures citations inside generated answers; an SEO audit measures positions on a results page. SEO optimizes a page to rank. GEO optimizes an entity to be quoted. The overlap is real (good content helps both), but the measurement, the tooling, and the fixes diverge sharply.

The clearest tell is in the data. When we pulled live search demand for the GEO category (DataForSEO, US/en, 2026-06-21), the buyer-intent terms were brand new and barely contested:

Term	Monthly volume	Keyword difficulty	CPC
geo audit	140	0	$15.01
ai visibility audit	90	6	$39.79
generative engine optimization services	590	6	$44.33
answer engine optimization	2,400	41	$25.24
generative engine optimization	5,400	57	$26.17

The category head terms ("generative engine optimization", "answer engine optimization") are already hard (KD 41 to 57). The buyer terms ("geo audit", "ai visibility audit") are wide open (KD 0 to 6) with $15 to $44 commercial CPCs. That gap is the whole point: demand for measuring AI visibility is arriving faster than the tools to do it. A GEO audit is the entry point. (For how the engines actually pick sources, see the companion breakdown on chudi.dev.)

How do you run a GEO audit in 7 steps?#

You run a GEO audit by listing your buyer questions, testing each one across the major AI engines, recording who gets cited, then checking your crawler access and content structure. Below is the exact sequence we use on paid audits, reproducible by hand.

Step 1: List your 20 real buyer questions#

Write the 20 questions a buyer actually types into ChatGPT before they buy in your category. Not keywords, full questions: "what is the best tool for X", "how do I do Y", "X vs Y". These are the queries your audit will test. If you cannot name 20, that is itself a finding.

Step 2: Test each question across the engines#

Run all 20 questions through ChatGPT, Perplexity, Gemini, and a Google AI Overview query. For each, record: were you cited (yes/no), what URL was cited, and which competitors appeared. Do this in a clean session with no personalization. This is the citation-rate core of the audit.

Step 3: Calculate your citation rate per engine#

Divide citations by total questions, per engine. A domain cited on 7 of 20 ChatGPT answers has a 35% ChatGPT citation rate. Track each engine separately, because they diverge hard: one of our own properties scored 50% on Perplexity and Claude but only 35% on ChatGPT, with the ChatGPT misses resolving to a mirror domain instead of the canonical site.

Step 4: Map who gets cited instead of you#

For every question where you were not cited, log the domains that were. Patterns emerge fast: usually 2 or 3 competitors own most of your category's AI answers. That list is your real competitive set inside AI search, and it is often different from your Google competitors.

Step 5: Check whether AI crawlers can read you#

Open your robots.txt and confirm you allow GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and Applebot-Extended. Then check for an llms.txt file at your root. If crawlers are blocked or there is no llms.txt, the engines may never see your best content no matter how good it is.

Step 6: Audit your content structure for liftability#

AI engines lift short, direct answers. Check whether your key pages open each section with a 40 to 60 word answer capsule, use question-led headings, and carry schema (FAQPage, HowTo, DefinedTerm, Organization). Walls of text without structure rarely get quoted.

Step 7: Rank the gaps and assign fixes#

Sort every gap by revenue impact, not by ease. A buyer query where a competitor is cited and you are absent outranks a cosmetic schema warning. The output is a prioritized fix list: crawler access first, then entity/schema, then answer-capsule rewrites, then new content for the queries you lose.

What tools do you need to run a GEO audit?#

You can run a basic GEO audit by hand with nothing but the four AI engines and your browser, plus a spreadsheet to log citations. The manual method is free and honest, and it is where every audit should start. The limit is repeatability: testing 20 questions across 4 engines monthly is tedious, and that is what dedicated AI-visibility tools automate.

The honest caveat: be wary of any tool that hands you a single composite "AI visibility score" with no breakdown. A real audit shows you the per-engine citation rate and the actual cited URLs, because a blended score hides exactly the engine where you are losing. Verifiable beats impressive.

How often should you re-run a GEO audit?#

Re-run a GEO audit monthly, because AI answers change far faster than Google rankings. Engines re-crawl, swap sources, and re-rank citations on a weekly-to-monthly cadence, so a one-time audit goes stale quickly. A baseline audit tells you where you stand; monthly monitoring tells you whether your fixes worked and whether a competitor just displaced you.

Run your GEO audit#

A GEO audit turns "are we visible in AI?" from a board-meeting shrug into a number you can act on. You can run the 7 steps above by hand today, or get a forensic AI Visibility Audit that tests your buyer queries across every engine, names the competitors cited in your place, and hands you the ranked fix list. Either way, the worst position is the one most brands are in right now: invisible inside AI search and not even measuring it.

Which AI engine do you think cites your brand the least right now, and have you ever actually checked?