Free Interactive Tool
AI SEO Scorecard - Will ChatGPT, Claude, and Gemini Cite You?
A complete GEO / AEO audit for getting cited by ChatGPT, Claude, Perplexity, Gemini, Copilot, and Google AI Overviews - weighted by what moves AI visibility most. Work top to bottom: access first, then structure, then content, then authority, then measurement. Needs Work earns half credit. Your progress saves automatically in this browser.
AI VisibilityJust getting started
0%
No leftover AI blocks in robots.txtA Disallow under GPTBot from 2023 is the single most common audit finding — check for old blanket blocks.
Search and citation bots explicitly allowedOAI-SearchBot, ChatGPT-User, ClaudeBot and Claude-SearchBot, PerplexityBot, Google-Extended, Applebot-Extended, meta-externalagent, CCBot, and Bingbot.
Training-bot policy decided separatelyTraining crawlers and search crawlers are different bots — you can block training while staying eligible for citations.
User-agents spelled exactly"GPT-Bot" instead of "GPTBot" silently fails — verify each spelling against the bot’s documentation.
AI allow rules above any wildcard Disallowrobots.txt is processed in order — allow rules below a wildcard block never fire.
CDN and firewall not blocking AI botsCloudflare’s defaults and bot-fight mode often block AI crawlers even when robots.txt allows them — check logs for 403s to GPTBot, ClaudeBot, PerplexityBot.
Server logs confirm AI bots are crawlingFilter access logs by user-agent and verify the bots actually reach your pages.
No login walls, paywalls, or aggressive interstitialsContent you want cited must be reachable without a login or dismissing overlays.
Content is server-side renderedMost AI crawlers don’t execute JavaScript — your words must be in the raw HTML (test with view-source or curl).
Clean semantic HTMLOne H1, logical H2/H3 hierarchy, real table, ul, and ol elements instead of styled divs.
Fast load timesTarget under ~2.5s LCP — slow pages get partial crawls.
Clean, descriptive URLsReadable slugs that describe the page, no parameter soup.
XML sitemap submitted and referenced in robots.txtBoth Google and Bing — Bing feeds ChatGPT and Copilot.
llms.txt at your domain rootA curated markdown summary pointing AI agents to your key pages — ~10% adoption makes it cheap differentiation.
Canonical tags set correctlyNo duplicate-content confusion across variants and parameters.
HTTPS everywhere and mobile-friendlyNo mixed content; Googlebot still feeds AI Overviews from the mobile index.
Broken links and redirect chains fixedAI crawlers abandon broken paths faster than Googlebot does.
Visible "Last updated" date on every pageAI systems weight freshness heavily and skip undated pages for time-sensitive queries.
Organization schema sitewideName, logo, sameAs links to all social profiles, founding date, and address.
Person schema for authorsName, credentials, job title, and sameAs to LinkedIn and other profiles.
Article schema on every postBlogPosting or Article with datePublished, dateModified, and author.
FAQPage and HowTo schemaOn Q&A pages and step-by-step guides respectively.
Product and Review schema on product pagesProduct plus Review or AggregateRating — Shopify handles much of this natively.
LocalBusiness schema if you serve an areaHours, service area, and phone matching your Google Business Profile.
BreadcrumbList and matching @typeSchema type matches what the page actually is — mismatches hurt trust.
Everything validatedGoogle’s Rich Results Test and the Schema.org validator both pass clean.
Answer first, explain secondOpen each page and major section with a direct 2–3 sentence answer (40–60 words), then supporting detail.
Question-phrased H2s and H3sHeadings mirror how people ask — "How much does X cost?" not "Pricing considerations."
One idea per paragraphShort paragraphs of 2–4 sentences chunk cleanly for retrieval.
Self-contained sectionsA passage lifted out of context still makes sense — exactly what RAG systems do with it.
Clear context statement in the openingWho this is for and what it covers, stated up front.
FAQ sections from real questionsSourced from support tickets, sales calls, and People Also Ask.
Tables and lists where they belongTables for comparisons, specs, and pricing; numbered lists for processes; bullets for options.
Terms defined, acronyms spelled outFirst use of every term and acronym is explained.
Written for query fan-out and topic clustersPages cover the sub-questions AI splits a query into, and pillar + supporting pages interlink so you own the whole topic.
Specific numbers, dates, and sources on every claimVerifiable facts get cited; vague claims get skipped.
Original dataSurveys, benchmarks, case studies, and proprietary research — the most cite-worthy asset you can create.
First-hand experience woven inReal screenshots, test results, "we ran this for 6 months and found…"
Named authors with bios and credentialsLinked profiles on every article.
Outbound citations and named expert quotesLink to authoritative references; quote experts with titles rather than "experts say."
Depth kept, thin content killedContent actually answers the question; outdated, thin, and near-duplicate pages consolidated or removed.
Citation-magnet content types in playComparison pages, buyer’s guides, real pricing pages, documentation, glossaries, calculators, benchmark reports.
Quarterly freshness programStats, pricing, screenshots, and dates reviewed at least quarterly — fresher pages outperform by up to 30% in AI visibility.
Identical brand info everywhereName, description, and NAP match across your site, Google Business Profile, LinkedIn, socials, and directories.
Consistent naming for products and peopleThe same names everywhere they appear.
A clear About pageWho you are, what you do, who you serve, and since when.
Wikipedia, Wikidata, and an active Google Business ProfileMajor training-data and AI Overview sources — pursue what you legitimately qualify for.
Brand-category association stated on-site"[Brand] is a [category] company that [what you do]" — LLMs learn these associations verbatim.
Present on your category’s review platformsG2, Capterra, TrustRadius for SaaS; Trustpilot and Google Reviews for consumer; TripAdvisor for travel.
Genuine Reddit participationOne of the most-cited sources across ChatGPT and Perplexity answers.
Quora answers and podcast appearancesHigh-intent questions in your niche; podcast transcripts are crawlable.
YouTube contentTranscripts get indexed and cited, and Gemini leans on YouTube.
Digital PR and industry publication mentionsMentions — not just links — teach models your brand exists.
Included in "best of" listicles and roundupsThese are the pages AI pulls from when someone asks "what’s the best X."
Backlink authority and active social profilesAI search seeds retrieval from conventional indexes, so domain authority still gates; LinkedIn and X corroborate the entity graph.
Indexed in Bing Webmaster ToolsBing powers ChatGPT search and Copilot — most brands forget it entirely.
You know each engine’s sourcesAI Overviews = Google index; ChatGPT = Bing + OAI-SearchBot; Perplexity = own crawler; Claude = Brave; Gemini = Google + YouTube.
Monthly brand-query tests on every platform"What is [brand]," "best [category]," "[brand] vs [competitor]," "[brand] pricing" — note what gets cited and what the AI gets wrong.
GA4 segments AI referral trafficchatgpt.com, perplexity.ai, claude.ai, gemini.google.com, copilot.microsoft.com — tracked to landing pages.
Search Console and Bing Webmaster monitoringAI Overviews and AI Mode visibility in GSC; indexing and Copilot visibility in Bing.
Server-log crawl monitoringWhich pages AI bots hit, and how often.
Mention rate and citation accuracy trackedTools like Profound, Otterly, Peec AI, or manual monthly prompt testing — and fix source content where the AI describes you wrong. Judge trends quarterly, not weekly.
Weights are directional, based on how AI retrieval systems work: crawler access gates everything (≈22%), then content structure (≈16%), technical foundation and content quality (≈14% each), structured data (≈12%), off-site corroboration (≈10%), entity consistency (≈7%), platform coverage and measurement (≈5%).
Want us to raise this score for you?We’ll audit your AI search visibility together and map the fastest wins — free, no strings.
Let’s Talk