A GEO Audit Checklist
A working GEO audit checklist across five areas: crawler access, content structure, substance, entities and measurement. Concrete yes-or-no checks to find the gaps.
A GEO audit checks whether your content and setup are actually built to be retrieved and cited by AI engines. It runs across four areas: access (can the crawlers reach you), structure (is your content chunkable and extractable), substance (does it carry the statistics, quotations and citations that win citation), and measurement (can you tell whether it’s working). This post is the working checklist — run it against any page or site to find the gaps between “ranks fine” and “gets cited.” Each item is a concrete yes-or-no you can check today.
Key takeaway
- A GEO audit covers four areas: crawler access, content structure, content substance, and measurement setup.
- Start with access — if AI search crawlers can’t reach you, no content work matters, and this is where sites most often fail silently.
- Work the checklist as concrete yes-or-no items; each “no” is a specific, fixable gap between ranking and being cited.
Area one: crawler access
Access comes first because it’s the prerequisite for everything else — an engine can’t cite a page it can’t fetch. Check each of these:
- Does robots.txt allow the AI search crawlers — OAI-SearchBot, PerplexityBot, Claude-SearchBot? Blocking these removes you from those engines’ answers.
- Is your CDN (Cloudflare or similar) silently blocking AI crawlers via an “AI scrapers” toggle that overrides robots.txt? Verify the setting, not just the file.
- Do your server logs actually show these crawlers reaching your pages? Confirmed access beats assumed access.
- Are your important pages indexed in Bing, not just Google? Bing feeds ChatGPT Search and Copilot, and it’s routinely neglected.
- Are pages fast and free of long redirect chains? Real-time search crawlers have low tolerance for slow pages and extra hops.
Area two: content structure
Structure decides whether your content chunks into clean, retrievable passages. Check each page for:
- Does each section open with a direct 40-to-60-word answer to its heading’s question, before any context?
- Are headings phrased as the actual questions buyers ask, not vague labels?
- Is every section self-contained — does it make sense read alone, with no orphan pronouns pointing at earlier sections?
- Are there clear, frequent headings so the page chunks along meaningful boundaries rather than splitting mid-thought?
- Is there a genuine FAQ block answering real People Also Ask questions in self-contained 40-to-60-word answers?
- Does any data trapped in an infographic also appear as HTML text, since engines can’t read images?
Passage-level
The level at which AI engines retrieve and cite — passages, not pages. Every structure check exists to make your individual passages cleanly extractable and self-contained, because a passage that can’t stand alone can’t be cited no matter how strong the surrounding page.
Source — GEO practice, grounded in Aggarwal et al., KDD 2024
Area three: content substance
Substance is what the KDD 2024 research showed actually drives citation. Audit each page for the material engines reach for:
- Does the page contain real statistics, each with a named source? Vague claims don’t get cited.
- Is there at least one relevant quotation from a credible source — the single strongest method in the research?
- Does the page cite and link to authoritative references for its claims?
- Is the content free of keyword stuffing, which scored below baseline and actively hurts?
- Are facts current and verified against primary sources, with honest publish or update dates?
- Is there a real, credentialed author byline linked to a substantive author page?
- Does the page genuinely cover the sub-questions of its topic, not just the head term?
Area four: entities and schema
This area makes your brand and content legible to entity-based retrieval. Check:
- Is Organization schema in place, with consistent name, details, and sameAs links to your profiles?
- Is Article and FAQPage schema present where relevant, and does it match the visible content exactly?
- Is your brand named consistently across your site, profiles, directories and review platforms?
- Are you present on the third-party sources AI engines lean on — review platforms, relevant communities, trade press?
Area five: measurement
Finally, confirm you can actually tell whether GEO is working. Without measurement, you’re optimising blind:
- Is there a fixed prompt set of 20 to 30 real buyer questions, run across engines on a monthly cadence?
- Is a GA4 AI-referral channel set up and ordered above Referral to isolate AI traffic?
- Are server logs being checked for AI crawler and real-time retrieval activity?
- Is Google Search Console’s AI reporting being reviewed for Google AI presence?
How to use the checklist
Run access first — a single “no” there can invalidate all other work, so fix it before anything else. Then work structure and substance page by page on your most important content, since those are where most citation gains live. Treat every “no” as a prioritised task: access issues are urgent, structure and substance are high-yield, schema and measurement are foundational. Re-run the audit quarterly, because the AI landscape and your content both change.
Frequently asked questions
What is a GEO audit?
A GEO audit assesses whether your content and technical setup are built to be retrieved and cited by AI engines. It covers four areas: crawler access (can engines reach you), content structure (is it chunkable and extractable), content substance (does it carry citable statistics, quotations and citations), and measurement (can you tell if it’s working). Each item is a concrete check that reveals gaps between ranking and being cited.
Where should a GEO audit start?
With crawler access, because an engine can’t cite a page it can’t reach. Confirm robots.txt allows the AI search crawlers, that your CDN isn’t silently blocking them, that server logs show them arriving, and that you’re indexed in Bing as well as Google. Access failures are common, silent, and invalidate all downstream content work — so rule them out first.
How often should I run a GEO audit?
Quarterly is a sensible cadence, plus a fresh check whenever you launch significant new content or change your hosting, CDN or site structure. The AI landscape shifts quickly — crawler names, engine behaviour and best practices all evolve — and your own content changes too. A quarterly re-audit catches new gaps before they quietly cost you citations.
What’s the most commonly failed GEO check?
Crawler access, usually via a CDN setting that blocks AI crawlers without the owner realising. Many sites also fail the structure checks — burying answers instead of leading with them, and writing sections that don’t stand alone. Access failures are the most damaging because they’re silent and total; structure failures are the most common source of “we rank but aren’t cited.”
The bottom line
A GEO audit turns “why aren’t we cited” into a concrete list of fixable gaps. Check access, structure, substance, entities and measurement, in that priority order, treating each “no” as a task. Access first because it’s total; structure and substance next because they’re where citation is won. Run it quarterly, and you’ll always know exactly where the daylight is between ranking and being the source AI engines quote.
We run this full GEO audit and fix every gap it surfaces as part of our AI Visibility service.
