# CitedMetrics — full content > AI visibility audits for B2B SaaS — data, not opinions. Full page contents follow; the compact index is at /llms.txt. --- # Home — AI Visibility Audit Source: https://citedmetrics.com/ # When buyers ask ChatGPT for tools like yours, who does it recommend? We run 50–150 buying-intent prompts through ChatGPT, Claude, Gemini and Perplexity to see whether they mention you, cite you, or recommend your competitors — data, not opinions. Get the audit — $950 USD Run a free snapshot first lock Delivered in 5 business days · No calls, no subscriptions · Secure checkout via Stripe 84% of B2B SaaS CMOs now use AI assistants to research vendors Wynter, 2026 ### Who AI recommends — gravity-fed water filters Based on 10 prompts × 2 engines · June 2026 Berkey 7 Sawyer 6 Alexapure 4 ProOne 3 Platypus 2 Your brand 0 mentions — invisible ## What is AI telling your customers right now? ChatGPT · June 2026 Exhibit A ### It recommends your competitors Buyer Best gravity-fed water filter for an off-grid cabin? smart_toy ChatGPT For off-grid living, the most recommended options are the Berkey Black Berkey system and the Sawyer Squeeze … Sources cited: homesteadfanatic.com thatyurt.com Not among them: you. Perplexity · June 2026 Exhibit B ### When it mentions a brand, it gets facts wrong Buyer Is [your brand] worth it? smart_toy Perplexity The company was acquired in 2023 , and entry models start at around $89 with a one-year warranty … error All three claims false — engines repeat stale sources until corrected. Illustrative example — audit findings carry raw-response evidence. Real engine output from a real product category, June 2026. One real brand in this category: zero mentions across all ten prompts. ## Why does this matter now? Your analytics stack was built for the search era. Buying moved to the answer era. -25% Drop in traditional search volume by 2026 as users shift to AI answers. Gartner 40%+ Of enterprise software buyers already use AI to research vendors. Industry Research ~80% Of the B2B buying cycle happens before they ever talk to sales. 6sense ## Not ready? See your own numbers first — free. A 1-page snapshot: 10 buying-intent prompts across ChatGPT and Perplexity — your mention rate vs your closest competitors. Send my snapshot Delivered to your inbox within 2 business days. No call, no pitch. Got it — your snapshot is queued. It lands in your inbox within 2 business days. ## What's in the audit? Up to 150 prompts × 4 engines × 3 runs — up to 1,800 scored answers behind every number. map ### Prompt coverage map 50–150 buying-intent prompts mapped to your funnel stages monitoring ### 4-engine visibility baseline Mentioned, cited, recommended, or invisible — per engine, per prompt pie_chart ### Competitor share-of-voice Who owns your category's answers, and by how much mood ### Sentiment & hallucination findings What AI gets wrong about you, documented with evidence troubleshoot ### Technical root-cause analysis AI-crawler access, schema, entity consistency, content structure, freshness route ### Prioritized 60–90 day fix roadmap Ranked by impact — built to hand to your team or agency Includes the raw data export of every engine output so your team can verify the findings. Not included: a retainer requirement, a sales pitch, or a dashboard to learn — the report is a diagnosis, and it's yours. ## What does the report look like? These are real pages from the sample audit — generated from live engine runs on a real product category, by the same pipeline that builds yours. Executive summary — verdict + KPIs Share of voice — all engines Explore the interactive dashboard → Read all 15 pages Yours covers your category, your competitors, your fixes — 50–150 prompts across 4 engines, every claim verified, with 90 days of dashboard access. ## How do we measure it? ### Audit specifications - Engines tested ChatGPT, Claude, Gemini, Perplexity — current production models, named with versions and run dates in your report. - Prompt volume 50–150 distinct buying-intent queries, branded and non-branded, mapped to your funnel. - Variance control Each prompt runs 3 times to average non-deterministic output — variance reported, not hidden. - Environment Clean sessions, no prior context, generic IP addresses. - Scoring Every answer coded mentioned, cited, recommended, or invisible against a fixed rubric. - Evidence Every finding links to the raw engine response, dated. ### How to judge any AI-visibility audit - check_circle Enough prompts to matter — 50+, not 5. - check_circle More than one engine — ChatGPT alone ignores where half your buyers ask. - check_circle Repeat runs — LLMs are stochastic; a single run is an anecdote, not data. - check_circle A competitor benchmark — your score means nothing without share-of-voice context. - check_circle Technical root causes — knowing you're invisible isn't enough; you need to know why. ## What are your alternatives? Option Price What you actually get | | Free AI checkers | $0 | A score. No evidence, no roadmap. | | Monitoring dashboards | ≈$337/mo, forever | You configure it, you interpret it. | | Enterprise AI-visibility platforms | $399–499/mo + sales call | Powerful — once you sit the demo and do the work. | | Thin productized audits | $497–1,497 | 5–20 prompts. | | Agency GEO retainers | $2,000–8,000/mo, 3-month minimum | Implementation included, diagnosis extra. | | CitedMetrics audit | $950, once | 50–150 prompts × 4 engines × 3 runs, report in 5 business days, no calls. | I built CitedMetrics to create the AI-visibility measurement I wished existed for my own 11 content sites. Before that I managed aerospace manufacturing at Rolls-Royce — overseeing the build and assembly of complete jet engine sections for Airbus and Boeing aircraft — where no decision was made without extensive data analysis. I apply the same standard here: every finding in your audit is backed by real prompts, real engine responses, and scored evidence. — David King, founder ## AI Visibility Audit $950 USD One-time fee. No subscription. Founding rate — rises to $1,500 once the first case studies publish. - check Prompt coverage map - check 4-engine visibility baseline - check Competitor share-of-voice - check Sentiment & hallucination findings - check Technical root-cause analysis - check Prioritized 60–90 day fix roadmap Get the audit Visa Mastercard Amex lock Secure checkout via Stripe No call. No account. No subscription. 0 Day 0 — Order + 5-min intake 1-4 Days 1–4 — We run 4 engines × 3 runs 5 Day 5 — Report in your inbox shield Read the full report. If it doesn't show you something about your AI visibility you didn't already know, reply to the delivery email within 14 days for a full refund — and keep the report. ## Frequently Asked Questions How is this different from SEO — or an SEO audit? expand_more AI engines synthesize answers from sources they trust instead of ranking ten links. We measure whether you're inside those answers and diagnose why not — content structure, entities, schema, AI-crawler access. It overlaps SEO; it isn't SEO. Which AI engines and models do you test? expand_more ChatGPT, Claude, Gemini and Perplexity — current production models, named with versions and run dates in your report. AI answers change constantly — how is this reliable? expand_more That's exactly why every prompt runs 3 times and variance is reported. A single run is an anecdote. An averaged, dated baseline is a measurement. What if my brand is already mentioned sometimes? expand_more Then the audit shows where you win, where you lose, to whom, and why. Most brands are partially visible — and unaware of which buying prompts they lose. What exactly arrives in 5 business days? expand_more A 20+ page PDF: visibility baseline, competitor share-of-voice, sentiment and hallucination findings with evidence, technical root-cause analysis, and a prioritized 60–90-day roadmap — plus the raw scored data. Do you implement the fixes too? expand_more Optionally, yes. A remediation retainer from $1,500/mo implements the roadmap and re-measures monthly — and your audit fee credits toward the first month. The audit also stands alone; many clients hand the roadmap to their own team. Why should I trust a new firm? expand_more You don't have to take our word: the methodology is published above, every finding links to raw engine output, and the keep-the-report guarantee means a useless report costs you nothing. Is my data confidential? expand_more Yes. Your report, prompts and results are never shared, resold, or used in our marketing without written permission. ## What does the research say? We publish our data and methods — the same evidence standard your audit gets. Research ### How B2B buyers use AI search in 2026 84% of B2B SaaS CMOs use LLMs in vendor discovery; 67% prefer rep-free buying; ~80% of shortlists form pre-contact — sourced numbers and what they mean. Research ### AI visibility audit & tool pricing — June 2026 market survey Original dataset: one-time audits $27–$4,500, monitoring tools $29–$654/mo (≈$337/mo average), enterprise platforms $399+/mo, agency retainers $2k–8k/mo. Guides ### How to get ChatGPT to recommend your business No paid placement exists — OpenAI confirms it. Recommendations come from evidence: crawler access, answer-shaped content, entity consistency, consensus, freshness. With June 2026 run data. All guides → All research → ## Find out what AI tells your buyers — before your competitor does. Get the audit — $950 USD Run a free snapshot first --- # About CitedMetrics Source: https://citedmetrics.com/about/ # Who is behind CitedMetrics? ## David King Founder I built CitedMetrics to create the AI-visibility measurement I wished existed for my own 11 content sites. Before that I managed aerospace manufacturing at Rolls-Royce — overseeing the build and assembly of complete jet engine sections for Airbus and Boeing aircraft — where no decision was made without extensive data analysis. I apply the same standard here: every finding in your audit is backed by real prompts, real engine responses, and scored evidence. LinkedIn → ## Why does CitedMetrics exist? Buyers moved their first question from Google to AI assistants, and most companies have no instrument that tells them what those assistants answer. CitedMetrics exists to be that instrument: real buying-intent prompts, four engines, repeated runs, scored evidence — delivered as a fixed-price audit instead of another dashboard subscription. ## How do we work? Async only, no sales calls. The methodology is published, every finding links to a raw engine response, and the keep-the-report guarantee means a useless report costs you nothing. Questions: support@citedmetrics.com . --- # Sample audit report Source: https://citedmetrics.com/sample-report/ # What does the audit report look like? Like this. Every page below was generated from real engine runs on a real product category — the same pipeline, charts, and scoring your report uses. The client name is illustrative; the data is not. Client editions cover your category with 50–150 prompts across 4 engines, verify every flagged claim, and include the raw data export. Explore the interactive dashboard Download the sample (PDF) Get yours — $950 USD Methodology, scoring rubric, and the evidence chain are documented inside the report and on the methodology page . --- # Guides Source: https://citedmetrics.com/guides/ # Guides Practical answers to the questions buyers and marketers actually ask about AI visibility — each one grounded in our own engine-run data. ### How to get ChatGPT to recommend your business No paid placement exists — OpenAI confirms it. Recommendations come from evidence: crawler access, answer-shaped content, entity consistency, consensus, freshness. With June 2026 run data. ### Why doesn't ChatGPT mention your company? Five checkable causes: blocked crawlers, no answer-shaped pages, thin third-party coverage, inconsistent entity facts, category confusion — and the diagnosis order. ### How to check what AI says about your brand The free 20-minute manual method and its three blind spots, then the measurement method: many prompts, repeated runs, clean sessions, competitor benchmarks. ### GEO vs SEO: what actually changed? SEO earns positions in links; GEO earns a place inside the answer. Same foundations, new scoreboard — and why neither replaces the other. ### AEO vs GEO: is there a real difference? AEO targets extracted answers (snippets, voice); GEO targets generated answers (ChatGPT, Perplexity). The work overlaps; the measurement differs. --- # How to get ChatGPT to recommend your business Source: https://citedmetrics.com/guides/how-to-get-chatgpt-to-recommend-your-business/ - Home - / Guides - / How do you get ChatGPT to recommend your business? # How do you get ChatGPT to recommend your business? Published 2026-06-10 · Updated 2026-06-10 · David King h2]:font-headline [&>h2]:text-2xl [&>h2]:font-bold [&>h2]:text-navy dark:[&>h2]:text-white [&>h2]:mt-10 [&>h2]:mb-3 [&>ul]:list-disc [&>ul]:pl-6 [&>ul]:space-y-2 [&_table]:w-full [&_table]:text-sm [&_table]:border-collapse [&_th]:text-left [&_th]:py-2 [&_th]:px-3 [&_th]:border-b-2 [&_th]:border-slate/20 dark:[&_th]:border-white/20 [&_th]:font-bold [&_th]:text-navy dark:[&_th]:text-white [&_td]:py-2 [&_td]:px-3 [&_td]:border-b [&_td]:border-slate/10 dark:[&_td]:border-white/10 [&_a]:text-teal dark:[&_a]:text-teal-bright [&_a]:underline"> TL;DR: you cannot pay your way into ChatGPT's answers — OpenAI says so, and our test runs confirm it. ChatGPT recommends a business when it finds consistent, credible evidence of that business, either in its training data or in a live web search. You can build that evidence deliberately. This guide shows how, using data from 88 logged engine runs we made in June 2026. ## Can you pay ChatGPT to recommend your business? No. There is no paid placement inside ChatGPT's organic answers. OpenAI's help center says answers are generated from the model's training and from search — not from advertising. The engines know this about themselves: when we asked the pay-to-play question across twelve runs in June 2026, they cited OpenAI's own help pages 40 times. Treat "guaranteed ChatGPT placement" offers the way you'd treat "guaranteed #1 on Google" — as a red flag. help.openai.com 48 youtube.com 23 semrush.com 15 ahrefs.com 9 community.openai.com 8 rankability.com 8 Domains the engines cited most across our 88 June 2026 runs on AI-visibility questions (ChatGPT gpt-5.5 + Perplexity sonar). OpenAI's own help pages dominate — followed by forums and video, which are exactly the sources a well-evidenced business can outrank. ## Where do ChatGPT's recommendations actually come from? Two places, and they behave differently. The first is the model's memory: patterns learned from training data. The second is live web search: pages fetched while answering. Our June 2026 runs made the split visible — GPT-5.5 answered several business questions with zero citations, straight from memory, while Perplexity cited sources on every single run. The same buyer question travels both paths. You need evidence on both. ## What makes a business recommendable? Five factors decide most of it. They are checkable, and they compound. - Crawler access. If GPTBot and OAI-SearchBot are blocked in your robots.txt, the search path cannot see you. Allow ClaudeBot and PerplexityBot while you're there — your buyers ask there too. - Answer-shaped content. Pages that answer a real buyer question in the first paragraph get quoted. Pages that warm up for four paragraphs get skipped. - Entity consistency. Same company name, same description, same facts — on your site, your LinkedIn, directories, everywhere. Conflicting facts make the model hedge or skip you. - Third-party consensus. Engines trust claims that repeat across independent sources. Reviews, comparisons, community threads, press — earned mentions, not just your own pages. - Freshness. Dated, current pages win citations. Stale pages lose them — and worse, engines keep repeating your old prices and dead products as if they were true. ## How do you win the live-search answers? This is the fast lane — results in weeks, not model-release cycles. Open your robots.txt to the AI crawlers. Publish pages that match real buyer questions, phrased the way buyers ask them. Put verifiable numbers and dates in them: engines prefer facts they can attribute. Add an llms.txt index so models can find your key pages. Our measurement methodology lists exactly what we check when we audit this. ## How do you get into the model's memory? Slower, but durable. Memory forms from training data, so it follows the consensus factor above: be described the same way in many credible places for long enough, and the model learns you as a fact. This lags by months — models retrain on cycles, not on your publishing schedule. Which is the strongest argument for starting now, not after your competitor becomes the answer. ## How do you know if it's working? Measure it the way you'd measure anything: repeatedly, against a baseline. Ask the questions your buyers ask, run each one several times (answers vary run to run — single checks mislead), and score whether you're mentioned, cited, recommended, or invisible. Benchmark your competitors with the same prompts. That's literally what our audit does across four engines — and the free 10-prompt snapshot is the small version. For why this channel decides deals before you ever see the buyer, see how B2B buyers use AI search in 2026 . ## Frequently asked questions - How do I make ChatGPT suggest my business? Make your business the best-evidenced answer. Allow OpenAI's crawlers, answer real buyer questions on your pages in the first paragraph, keep your company facts identical everywhere, and earn mentions on credible third-party pages. Then measure monthly with repeated prompts — single checks are anecdotes. - Can you train ChatGPT to recommend your business? Not the public one. You cannot submit your business to the model or fine-tune what it tells other people. A custom GPT only changes answers inside that custom GPT. The public model follows evidence: training-data presence plus what web search finds at answer time. - How do I get my business ranked on ChatGPT? There is no ranking to climb — each answer is generated fresh, and the same question can return different names on different runs. What you influence is the probability of being named. That probability follows crawler access, answer-shaped content, entity consistency, third-party consensus, and freshness. - Can you pay ChatGPT to recommend your business? No. OpenAI states that answers are not sponsored and placement is not for sale. In our June 2026 test runs, the engines themselves cited OpenAI's help center 40 times when asked this question. Anyone selling guaranteed ChatGPT placement is selling something they do not control. ## Keep reading Guides Why doesn't ChatGPT mention your company? Guides How to check what AI says about your brand Research CitedMetrics audit methodology Written by David King · Founder I built CitedMetrics to create the AI-visibility measurement I wished existed for my own 11 content sites. Before that I managed aerospace manufacturing at Rolls-Royce — overseeing the build and assembly of complete jet engine sections for Airbus and Boeing aircraft — where no decision was made without extensive data analysis. I apply the same standard here: every finding in your audit is backed by real prompts, real engine responses, and scored evidence. --- # Why doesn't ChatGPT mention your company? Source: https://citedmetrics.com/guides/why-chatgpt-doesnt-mention-your-company/ - Home - / Guides - / Why doesn't ChatGPT mention your company? # Why doesn't ChatGPT mention your company? Published 2026-06-10 · Updated 2026-06-10 · David King h2]:font-headline [&>h2]:text-2xl [&>h2]:font-bold [&>h2]:text-navy dark:[&>h2]:text-white [&>h2]:mt-10 [&>h2]:mb-3 [&>ul]:list-disc [&>ul]:pl-6 [&>ul]:space-y-2 [&_table]:w-full [&_table]:text-sm [&_table]:border-collapse [&_th]:text-left [&_th]:py-2 [&_th]:px-3 [&_th]:border-b-2 [&_th]:border-slate/20 dark:[&_th]:border-white/20 [&_th]:font-bold [&_th]:text-navy dark:[&_th]:text-white [&_td]:py-2 [&_td]:px-3 [&_td]:border-b [&_td]:border-slate/10 dark:[&_td]:border-white/10 [&_a]:text-teal dark:[&_a]:text-teal-bright [&_a]:underline"> TL;DR: ChatGPT isn't ignoring you — it has no evidence of you, or it can't read the evidence you have. The cause is almost always one of five checkable things: blocked AI crawlers, pages that don't answer buyer questions, thin third-party coverage, inconsistent company facts, or a category the model doesn't connect you to. This guide gives you the diagnosis order. It is the same one our audit runs. ## Is your site blocked at the front door? Check robots.txt first. Thousands of sites blocked GPTBot in 2023–2024 — often via a CDN toggle nobody remembers — and that block now mutes them in AI answers. Look for rules against GPTBot, OAI-SearchBot, ClaudeBot and PerplexityBot. Some firewalls also refuse automated requests entirely: in our June 2026 category test, one major brand's site was unreachable to any non-browser agent. That blocks the crawlers that feed answers too. ## Do you have pages that answer buyer questions? Engines quote pages that answer questions directly. If your site has product pages and a blog but nothing that plainly answers "best X for Y" or "is X worth it", there is nothing for the search path to lift. Brands that win recommendations have answer-shaped pages for the exact questions buyers ask — phrased the way buyers ask them. ## Does anyone else talk about you? This is the big one. Engines trust claims that repeat across independent sources. In our June 2026 runs, the sources behind recommendation answers were review sites, comparison posts and community threads — not the brands' own sites. If your third-party footprint is thin, the model has one voice saying you exist: yours. One voice is not consensus, and consensus is what gets recommended. ## Are your facts consistent everywhere? Different descriptions on your site, LinkedIn, directories and old press releases make the model hedge. Entity confusion is worse than absence: the engine may merge you with a similarly named company, repeat stale pricing, or describe a product you killed years ago. Same name, same description, same facts — everywhere. ## Does the model know what category you're in? Models answer category questions from category associations. If nothing durable links your brand to "gravity water filters" or "CRM for agencies", you won't surface when buyers ask category questions — even if the model knows your name. Category-anchored content and category-anchored third-party mentions build that link. ## How do you diagnose it properly? In this order: crawler access, answer-shaped pages, third-party coverage, entity consistency, category linkage. Each has a different fix, and guessing wastes quarters. Measurement settles it: ask the real buying questions repeatedly across engines and score the answers — our methodology shows the rubric. The free snapshot is the quick version; the full audit traces every zero to its cause and hands you the fix order. If you want the mechanics of becoming recommendable, read how to get ChatGPT to recommend your business . ## Frequently asked questions - How do I make my company appear on ChatGPT? Fix the evidence chain: allow AI crawlers in robots.txt, publish pages that answer real buyer questions directly, keep your company facts identical across the web, and earn mentions on the third-party sites engines already cite in your category. Then re-test with repeated prompts. - How do you add your company to ChatGPT? You can't submit it anywhere — there is no index to join and no form to fill in. ChatGPT learns about companies from training data and live web search. Both are influenced by publishing, consistency, and third-party coverage, not by registration. - Why did my company block ChatGPT? Many sites blocked GPTBot in robots.txt during 2023–2024, often by default in a CDN or plugin setting. That choice now silences the site in AI answers. Check your robots.txt for GPTBot, OAI-SearchBot, ClaudeBot and PerplexityBot rules before assuming a content problem. - How does your business show up on ChatGPT? Through two paths: the model's training data (slow, built from durable third-party consensus) and live web search (fast, built from crawlable answer-shaped pages). A business that covers both paths shows up; a business that covers neither stays invisible. ## Keep reading Guides How to get ChatGPT to recommend your business Guides How to check what AI says about your brand Research What is an AI search visibility audit? Written by David King · Founder I built CitedMetrics to create the AI-visibility measurement I wished existed for my own 11 content sites. Before that I managed aerospace manufacturing at Rolls-Royce — overseeing the build and assembly of complete jet engine sections for Airbus and Boeing aircraft — where no decision was made without extensive data analysis. I apply the same standard here: every finding in your audit is backed by real prompts, real engine responses, and scored evidence. --- # How to check what AI says about your brand Source: https://citedmetrics.com/guides/how-to-check-what-ai-says-about-your-brand/ - Home - / Guides - / How do you check what AI says about your brand? # How do you check what AI says about your brand? Published 2026-06-10 · Updated 2026-06-10 · David King h2]:font-headline [&>h2]:text-2xl [&>h2]:font-bold [&>h2]:text-navy dark:[&>h2]:text-white [&>h2]:mt-10 [&>h2]:mb-3 [&>ul]:list-disc [&>ul]:pl-6 [&>ul]:space-y-2 [&_table]:w-full [&_table]:text-sm [&_table]:border-collapse [&_th]:text-left [&_th]:py-2 [&_th]:px-3 [&_th]:border-b-2 [&_th]:border-slate/20 dark:[&_th]:border-white/20 [&_th]:font-bold [&_th]:text-navy dark:[&_th]:text-white [&_td]:py-2 [&_td]:px-3 [&_td]:border-b [&_td]:border-slate/10 dark:[&_td]:border-white/10 [&_a]:text-teal dark:[&_a]:text-teal-bright [&_a]:underline"> TL;DR: there are two ways. The manual method is free and takes 20 minutes: ask the engines your buyers' questions and read the answers. It is worth doing today — and it has three blind spots that make it unreliable as a measurement. The measurement method fixes them: many prompts, repeated runs, clean sessions, and scoring against competitors. Use the first to get alarmed, the second to get answers. ## What is the manual method? Open ChatGPT and Perplexity. Ask the questions a buyer would ask: "best [your category] for [your buyer]", "[your brand] vs [competitor]", "is [your brand] worth it", "[competitor] alternatives". Read what comes back. Note who gets named, who gets recommended, and what the engines claim about you. Twenty minutes of this is more honest than most brand decks. Two hygiene rules while you do it: use a private window or an account that hasn't researched your own brand, so personalization doesn't flatter you. And copy answers out verbatim with the date — engine answers change, and an undated screenshot proves nothing in next month's argument. ## Where does the manual method mislead you? - Personalization. Your logged-in session knows you. It may name your brand because you keep asking about it. Your buyers' sessions won't. - Single runs. Engines vary between runs on the same question. In our June 2026 testing we ran every prompt three times because one-run results kept disagreeing with themselves. - Coverage. You'll ask five questions. Buyers ask across the whole funnel — comparison, validation, problem, alternatives. The question you skip is often the one you lose. ## Which questions should you actually ask? Cover the four families buyers use. Swap in your own category and rivals: - Comparison: "best [category] for [your buyer type]" · "top [category] tools in 2026" - Validation: "is [your brand] worth it" · "[your brand] reviews — what do customers say" - Problem: "how do I solve [the problem you fix]" — does the answer route to your category at all? - Alternatives: "[competitor] alternatives" · "[your brand] vs [competitor]" Record three things per answer: who got named, who got recommended, and any factual claim about you — price, products, ownership. A plain spreadsheet beats memory: one row per question, one column per engine, dated. ## What does a real measurement look like? Four properties, none optional: enough prompts to matter (50+, mapped to funnel stages), more than one engine, repeated runs with variance reported, and competitor benchmarking — your score means nothing without share-of-voice context. Score every answer against a fixed rubric: mentioned, cited, recommended, or invisible. That's the entire published methodology ; demand the same from anyone you hire. ## What should you do with the results? Three outcomes, three moves. Invisible: diagnose why — start with the five checkable causes . Mentioned but misdescribed: hunt the stale sources the engines repeat, fix them at the origin. Visible and accurate: defend it — re-measure monthly, because engines change weekly and competitors are working on the same answers. ## What's the fastest way to start? The free snapshot : 10 buying-intent prompts across ChatGPT and Perplexity, your mention rate vs your closest competitors, in your inbox within 2 business days. The full audit is the measurement-grade version — 50–150 prompts, 4 engines, 3 runs each, with the report to show for it. ## Frequently asked questions - Can I just ask ChatGPT what it thinks of my brand? Yes — and you should, today. But treat one answer as an anecdote, not a finding. Engines vary run to run, your logged-in session is personalized, and a single prompt covers one question of the dozens buyers actually ask. A real check uses many prompts, repeated runs, and a clean environment. - Which AI is best for checking brand sentiment? Check the engines your buyers use, not the one that flatters you. In practice that means ChatGPT and Perplexity first, then Gemini and Claude. They behave differently: in our June 2026 runs, ChatGPT often answered from memory with no citations, while Perplexity cited sources on every single run. - How often should I re-check? Monthly, with the same prompt set, so changes are comparable. Engines update continuously; a check older than a quarter describes an engine that no longer exists. ## Keep reading Guides Why doesn't ChatGPT mention your company? Research CitedMetrics audit methodology Research How B2B buyers use AI search in 2026 Written by David King · Founder I built CitedMetrics to create the AI-visibility measurement I wished existed for my own 11 content sites. Before that I managed aerospace manufacturing at Rolls-Royce — overseeing the build and assembly of complete jet engine sections for Airbus and Boeing aircraft — where no decision was made without extensive data analysis. I apply the same standard here: every finding in your audit is backed by real prompts, real engine responses, and scored evidence. --- # GEO vs SEO Source: https://citedmetrics.com/guides/geo-vs-seo/ - Home - / Guides - / GEO vs SEO: what's actually different? # GEO vs SEO: what's actually different? Published 2026-06-10 · Updated 2026-06-10 · David King h2]:font-headline [&>h2]:text-2xl [&>h2]:font-bold [&>h2]:text-navy dark:[&>h2]:text-white [&>h2]:mt-10 [&>h2]:mb-3 [&>ul]:list-disc [&>ul]:pl-6 [&>ul]:space-y-2 [&_table]:w-full [&_table]:text-sm [&_table]:border-collapse [&_th]:text-left [&_th]:py-2 [&_th]:px-3 [&_th]:border-b-2 [&_th]:border-slate/20 dark:[&_th]:border-white/20 [&_th]:font-bold [&_th]:text-navy dark:[&_th]:text-white [&_td]:py-2 [&_td]:px-3 [&_td]:border-b [&_td]:border-slate/10 dark:[&_td]:border-white/10 [&_a]:text-teal dark:[&_a]:text-teal-bright [&_a]:underline"> TL;DR: SEO earns positions in a list of links; GEO (generative engine optimization) earns a place inside the answer itself. The foundations overlap — crawlability, structure, authority — but the reader changed: a model now decides what to repeat to your buyer. GEO is not replacing SEO. It is replacing the scoreboard. ## What does each one optimize for? SEO GEO | | What gets optimized | Pages for ranked lists of links | Evidence for synthesized answers | | Unit of success | Position #1–10, clicks | Mentioned, cited, recommended | | Query shape | Keywords | Full questions in buyer language | | Who reads first | A person scanning results | A model choosing what to repeat | | Key inputs | Links, on-page signals, authority | Crawler access, answer-shaped content, entity consistency, third-party consensus, freshness | | Scoreboard | Rank trackers, traffic | Share of voice across repeated prompt runs | ## Why does the difference matter to a buyer journey? In search, ten results compete for a click and the buyer reads the source. In an answer, the engine has already chosen two or three names — and most buyers never see the sources. The competition happens before the click, inside the synthesis. That is why a brand can rank well and still be invisible: ranking is an input, not the verdict. The verdict is whether the model repeats you. ## What carries over from SEO — and what doesn't? Carries over: technical health, crawlable architecture, genuine authority, content that answers real questions. Doesn't: chasing keywords instead of questions, optimizing for CTR instead of quotability, and reading rank trackers as the scoreboard. The new inputs that matter most — AI-crawler access, entity consistency, third-party consensus — barely appear in classic SEO checklists, and they decide who engines recommend . ## What are the common mistakes when teams add GEO? - Renaming the SEO report. Calling rankings "AI visibility" measures nothing new. The answer side needs its own instrument: repeated prompts, scored states. - Blocking the crawlers, keeping the dashboards. Plenty of sites still block GPTBot from a 2023 default while investing in content the engines are never allowed to read. - Writing for extraction, forgetting consensus. One perfect page can win a snippet; recommendations need third-party corroboration too. - One-off checks. A single ChatGPT screenshot ages in weeks. Engines change continuously; only a re-measured baseline shows direction. ## Is GEO replacing SEO? No. Engines feed on the open web that SEO disciplines built, so gutting SEO starves GEO. But the share of buying decisions formed inside answers keeps growing — 84% of B2B SaaS CMOs now use LLMs in vendor research (Wynter, 2026). Run both; score both. Most teams' real gap is that they measure rankings weekly and answers never. ## How do you know where you stand? Measure the answer side the way you already measure rankings: repeated prompts, multiple engines, scored against competitors — the methodology is published. The free snapshot gives you the first reading; the audit gives you the diagnosis and the fix order. ## Frequently asked questions - Is GEO replacing SEO? No — it sits on top of it. AI engines lean on crawlable, well-structured, authoritative content, which is what SEO builds. But ranking is no longer the finish line: being quoted inside the answer is. Teams that treat GEO as a replacement skip the foundations; teams that treat SEO as sufficient become invisible in answers. - Why is AI SEO called GEO? GEO stands for generative engine optimization — the term comes from a 2023 academic paper (Aggarwal et al., arXiv) that measured how content changes affect inclusion in generated answers. You will also see AEO (answer engine optimization) and LLM SEO for overlapping ideas. - Which is better, SEO or GEO? Wrong question — they pay in different currencies. SEO pays in ranked links and clicks. GEO pays in mentions, citations and recommendations inside answers, where a growing share of B2B buying decisions form. Most companies need both; which to fund first depends on where their buyers actually ask. - What's replacing SEO? Nothing is replacing the work — crawlability, structure, authority all still matter. What is being replaced is the scoreboard: from rankings and traffic toward share of voice inside AI answers. Measure both and the rest of the debate resolves itself. ## Keep reading Guides AEO vs GEO: is there a real difference? Guides How to get ChatGPT to recommend your business Research How B2B buyers use AI search in 2026 Written by David King · Founder I built CitedMetrics to create the AI-visibility measurement I wished existed for my own 11 content sites. Before that I managed aerospace manufacturing at Rolls-Royce — overseeing the build and assembly of complete jet engine sections for Airbus and Boeing aircraft — where no decision was made without extensive data analysis. I apply the same standard here: every finding in your audit is backed by real prompts, real engine responses, and scored evidence. --- # AEO vs GEO Source: https://citedmetrics.com/guides/aeo-vs-geo/ - Home - / Guides - / AEO vs GEO: is there a real difference? # AEO vs GEO: is there a real difference? Published 2026-06-10 · Updated 2026-06-10 · David King h2]:font-headline [&>h2]:text-2xl [&>h2]:font-bold [&>h2]:text-navy dark:[&>h2]:text-white [&>h2]:mt-10 [&>h2]:mb-3 [&>ul]:list-disc [&>ul]:pl-6 [&>ul]:space-y-2 [&_table]:w-full [&_table]:text-sm [&_table]:border-collapse [&_th]:text-left [&_th]:py-2 [&_th]:px-3 [&_th]:border-b-2 [&_th]:border-slate/20 dark:[&_th]:border-white/20 [&_th]:font-bold [&_th]:text-navy dark:[&_th]:text-white [&_td]:py-2 [&_td]:px-3 [&_td]:border-b [&_td]:border-slate/10 dark:[&_td]:border-white/10 [&_a]:text-teal dark:[&_a]:text-teal-bright [&_a]:underline"> TL;DR: AEO (answer engine optimization) grew up around extracted answers — featured snippets, People-Also-Ask, voice assistants — where an engine lifts one passage and shows it. GEO (generative engine optimization) targets generated answers — ChatGPT, Perplexity, AI Overviews — where a model writes a synthesis and decides who gets named. The work overlaps almost entirely. The difference that matters is how you measure. ## What did AEO mean? Win the answer box. Structure a page so an engine can extract a clean, complete passage: the question as a heading, the answer in the first sentence, schema marking it up. One page, one extraction, one winner — and you could see the win in the search results. ## What does GEO add? Generated answers aren't extracted from one page — they're synthesized from many sources, filtered by what the model trusts. So GEO adds the evidence layer: AI-crawler access, entity consistency, consensus across third-party sources, freshness. One perfect page can win a snippet; only a body of corroborated evidence wins a recommendation. That's also why the original GEO paper measured inclusion in answers , not positions. ## Where do they overlap? Almost everywhere it counts: answer the question directly, in the buyer's words, near the top of the page; keep structure clean and schema honest; be the source engines can quote without editing. A page built to win a snippet is usually a page a generative engine can cite. Do the work once, then check both scoreboards. ## Which surfaces does each label cover? AEO's home turf: Google featured snippets, People-Also-Ask boxes, voice assistants reading one answer aloud, and on-site answer widgets. GEO's turf: ChatGPT, Perplexity, Claude, Gemini, Google's AI Overviews and AI Mode, Copilot — anywhere a model writes the answer instead of quoting one passage. The surfaces are converging: AI Overviews sit on a search page but synthesize like a chat engine. That convergence is why the labels blur and the work doesn't. The stakes converged too. In our June 2026 SERP checks, nearly every commercial question we tested in this space carried an AI Overview above the classic results. The "answer surface" is no longer a niche feature you optimize for separately — it is the first thing a buyer reads on most queries that matter. ## What should a team actually do on Monday? - Pick the buyer questions that matter — comparison, validation, problem, alternatives — and write pages that answer each one in the first paragraph. - Mark them up with honest schema and keep your company facts identical everywhere. - Open robots.txt to the AI crawlers and publish an llms.txt index. - Earn third-party coverage on the sources engines already cite in your category. - Then check both scoreboards: snippet wins in search results, mention and recommendation rates in repeated engine runs. ## How does the measurement differ? AEO is checkable in a search result: you hold the snippet or you don't. Generated answers vary run to run and engine to engine, so GEO needs repeated runs, multiple engines, and scoring — mentioned, cited, recommended, invisible — against competitors. That measurement discipline is the methodology behind our audit , and the difference between knowing your status and guessing it. For the wider context war, see GEO vs SEO . ## Frequently asked questions - Do I need both AEO and GEO? In practice you do one body of work and both labels claim it: direct answers to real questions, clean structure and schema, consistent entity facts, credible third-party coverage. Pick either label internally; just measure results across both surfaces — featured answers and generated answers. - Where did the term GEO come from? From a 2023 academic paper (Aggarwal et al., "GEO: Generative Engine Optimization", arXiv) that tested which content changes increase inclusion in generated answers. AEO is older, coined for answer boxes and voice assistants before chat engines took over. - Is AEO dead now that everyone says GEO? The label is fading; the work is not. Featured snippets, People-Also-Ask boxes and voice answers still exist and still convert. GEO simply extended the same discipline to engines that write whole answers instead of quoting one. ## Keep reading Guides GEO vs SEO: what actually changed? Research CitedMetrics audit methodology Research What is an AI search visibility audit? Written by David King · Founder I built CitedMetrics to create the AI-visibility measurement I wished existed for my own 11 content sites. Before that I managed aerospace manufacturing at Rolls-Royce — overseeing the build and assembly of complete jet engine sections for Airbus and Boeing aircraft — where no decision was made without extensive data analysis. I apply the same standard here: every finding in your audit is backed by real prompts, real engine responses, and scored evidence. --- # Research Source: https://citedmetrics.com/research/ # Research Original data and published methodology. Everything here is dated, sourced, and written to be checked. ### How B2B buyers use AI search in 2026 84% of B2B SaaS CMOs use LLMs in vendor discovery; 67% prefer rep-free buying; ~80% of shortlists form pre-contact — sourced numbers and what they mean. ### CitedMetrics audit methodology Prompt taxonomy, 4 engines, 3-run variance control, mentioned/cited/recommended/invisible scoring rubric, evidence chain. ### AI visibility audit & tool pricing — June 2026 market survey Original dataset: one-time audits $27–$4,500, monitoring tools $29–$654/mo (≈$337/mo average), enterprise platforms $399+/mo, agency retainers $2k–8k/mo. ### What is an AI search visibility audit? A diagnostic measuring whether AI engines mention, cite, or recommend a brand: repeated prompt runs, fixed scoring rubric, competitor benchmarks, fix roadmap. --- # How B2B buyers use AI search in 2026 Source: https://citedmetrics.com/research/ai-search-buyer-behavior-2026/ - Home - / Research - / How do B2B buyers actually use AI search in 2026? # How do B2B buyers actually use AI search in 2026? Published 2026-06-10 · Updated 2026-06-10 · David King h2]:font-headline [&>h2]:text-2xl [&>h2]:font-bold [&>h2]:text-navy dark:[&>h2]:text-white [&>h2]:mt-10 [&>h2]:mb-3 [&>ul]:list-disc [&>ul]:pl-6 [&>ul]:space-y-2 [&_table]:w-full [&_table]:text-sm [&_table]:border-collapse [&_th]:text-left [&_th]:py-2 [&_th]:px-3 [&_th]:border-b-2 [&_th]:border-slate/20 dark:[&_th]:border-white/20 [&_th]:font-bold [&_th]:text-navy dark:[&_th]:text-white [&_td]:py-2 [&_td]:px-3 [&_td]:border-b [&_td]:border-slate/10 dark:[&_td]:border-white/10 [&_a]:text-teal dark:[&_a]:text-teal-bright [&_a]:underline"> TL;DR: the people who buy B2B software now ask AI assistants first. 84% of B2B SaaS CMOs use LLMs in vendor discovery (Wynter, 2026 — up from 24% a year earlier), 67% of B2B buyers prefer a rep-free buying experience (Gartner, 2026), and ~80% of vendor shortlists are formed before any sales contact (6sense). If AI answers shape the shortlist and the shortlist decides the deal, AI answers are a revenue input — currently unmeasured at most companies. CMOs using LLMs, 2026 84% CMOs using LLMs, 2025 24% Prefer rep-free buying 67% Shortlist set pre-contact ~80% AI-assisted B2B buying, 2025–2026. Sources: Wynter CMO panel 2026 (LLM adoption), Gartner 2026 (rep-free preference), 6sense 2025 (pre-contact shortlists). ## What do the numbers say? Finding Number Source, year | | B2B SaaS CMOs using LLMs in vendor discovery | 84% (24% in 2025) | Wynter, 2026 | | B2B buyers preferring a rep-free buying experience | 67% | Gartner, 2026 | | Buying time spent with any vendor's sales team | ~17% | Gartner | | Vendor shortlist formed on day one | ~80% | 6sense, 2025 | | Deals won by the pre-contact favorite | 77–80% | 6sense, 2025 | | Decision-makers moved to research something by educational content | 75% | Edelman/LinkedIn, 2024 | ## What does this mean for a B2B SaaS brand? - The first sales conversation happens without you. If a CMO's assistant of choice answers "best tools for X" without naming you, you are not losing the deal late — you were never in it. - Search dashboards don't see it. Rankings and traffic reports measure the search era. Answer-era visibility — whether engines mention, cite, or recommend you — has no default dashboard. - It is measurable. Run real buying-intent prompts across the major engines, repeat the runs to control for variance, and score the answers. That is the entire premise of our methodology . ## Why doesn't your analytics stack show any of this? Search consoles and rank trackers were built to watch links, not answers. When a buyer asks an assistant "best tools for X" and acts on the reply, no referrer fires, no impression logs, and no dashboard moves. The visit you eventually see — if you see one — looks like direct traffic. The influence happened upstream, inside the answer. That is why teams with mature SEO reporting can still be blind here: the instrument measures the old channel. ## How should you respond first? Measure before you optimize. A dated, multi-engine baseline tells you whether you have an exposure problem, a citation problem, or a hallucination problem — three different fixes. The CitedMetrics audit is one way to get that baseline; the methodology link above shows how to demand the same rigor from anyone you hire. Whatever you do, re-measure on a schedule: engines change weekly, and a one-off check ages fast. ## Keep reading Guides How to check what AI says about your brand Guides GEO vs SEO: what actually changed? Research What is an AI search visibility audit? Written by David King · Founder I built CitedMetrics to create the AI-visibility measurement I wished existed for my own 11 content sites. Before that I managed aerospace manufacturing at Rolls-Royce — overseeing the build and assembly of complete jet engine sections for Airbus and Boeing aircraft — where no decision was made without extensive data analysis. I apply the same standard here: every finding in your audit is backed by real prompts, real engine responses, and scored evidence. --- # Audit methodology Source: https://citedmetrics.com/research/methodology/ - Home - / Research - / How we measure AI visibility: full methodology # How we measure AI visibility: full methodology Published 2026-06-10 · Updated 2026-06-10 · David King h2]:font-headline [&>h2]:text-2xl [&>h2]:font-bold [&>h2]:text-navy dark:[&>h2]:text-white [&>h2]:mt-10 [&>h2]:mb-3 [&>ul]:list-disc [&>ul]:pl-6 [&>ul]:space-y-2 [&_table]:w-full [&_table]:text-sm [&_table]:border-collapse [&_th]:text-left [&_th]:py-2 [&_th]:px-3 [&_th]:border-b-2 [&_th]:border-slate/20 dark:[&_th]:border-white/20 [&_th]:font-bold [&_th]:text-navy dark:[&_th]:text-white [&_td]:py-2 [&_td]:px-3 [&_td]:border-b [&_td]:border-slate/10 dark:[&_td]:border-white/10 [&_a]:text-teal dark:[&_a]:text-teal-bright [&_a]:underline"> TL;DR: we measure what AI engines tell buyers about a brand using 50–150 real buying-intent prompts, run across ChatGPT, Claude, Gemini and Perplexity, three times each, scored against a fixed rubric, with every finding linked to a dated raw response. This page publishes the method in full — both so audit buyers can verify what they paid for, and so anyone evaluating any AI-visibility vendor knows what to demand. ## How are prompts selected? Prompts are generated from the client's category, not from their keywords: comparison questions ("best X for Y"), problem questions ("how do I solve Z"), validation questions ("is [brand] worth it"), and alternative-seeking questions ("[competitor] alternatives") — branded and non-branded, mapped to funnel stages. 50–150 distinct prompts per audit, disclosed in the report's prompt appendix. ## How is each answer scored? - Mentioned — the brand appears anywhere in the answer. - Cited — the brand's own site (or a page about it) is used as a source. - Recommended — the engine names the brand as a pick, not just a mention. - Invisible — none of the above. The zero rows are usually the reason an audit gets commissioned. Sentiment and factual accuracy are coded separately: what the engine says about the brand, and whether it's true. False claims (discontinued products, wrong pricing, phantom acquisitions) are flagged as hallucination findings with the raw response attached. ## How is engine randomness handled? Every prompt runs three times per engine in clean sessions — no prior context, generic IPs. Scores are averaged and the variance is reported, not hidden. High-variance prompts are flagged: an answer that names you once in three runs is a coin-flip, not a presence. ## What's in the evidence chain? Every number in the report traces to raw engine responses with model versions and run dates. The raw export ships with the report so a client's team can re-check any claim. We consider this table stakes; the market mostly doesn't provide it. ## How is AI visibility calculated? Four rates, all over repeated runs. Mention rate: runs naming the brand ÷ total runs. Citation rate: runs where the brand's own site is a source ÷ total runs. Recommendation rate: runs endorsing the brand ÷ total runs. Share of voice: the brand's mentions ÷ all brand mentions in the category. Every rate is reported per engine and overall, with the run count attached — a rate without its denominator is marketing, not measurement. ## How should you judge any AI-visibility audit (including ours)? - Enough prompts to matter — 50+, not 5. - More than one engine — ChatGPT alone ignores where half your buyers ask. - Repeat runs — single runs are anecdotes. - A competitor benchmark — your score means nothing without share-of-voice context. - Technical root causes — knowing you're invisible isn't enough; you need to know why. ## Keep reading Research What is an AI search visibility audit? Guides How to get ChatGPT to recommend your business Research AI visibility audit & tool pricing — June 2026 market survey Written by David King · Founder I built CitedMetrics to create the AI-visibility measurement I wished existed for my own 11 content sites. Before that I managed aerospace manufacturing at Rolls-Royce — overseeing the build and assembly of complete jet engine sections for Airbus and Boeing aircraft — where no decision was made without extensive data analysis. I apply the same standard here: every finding in your audit is backed by real prompts, real engine responses, and scored evidence. --- # AI visibility audit pricing — June 2026 survey Source: https://citedmetrics.com/research/ai-visibility-audit-pricing-market/ - Home - / Research - / AI visibility audit pricing — June 2026 survey # AI visibility audit pricing — June 2026 survey Published 2026-06-10 · Updated 2026-06-10 · David King h2]:font-headline [&>h2]:text-2xl [&>h2]:font-bold [&>h2]:text-navy dark:[&>h2]:text-white [&>h2]:mt-10 [&>h2]:mb-3 [&>ul]:list-disc [&>ul]:pl-6 [&>ul]:space-y-2 [&_table]:w-full [&_table]:text-sm [&_table]:border-collapse [&_th]:text-left [&_th]:py-2 [&_th]:px-3 [&_th]:border-b-2 [&_th]:border-slate/20 dark:[&_th]:border-white/20 [&_th]:font-bold [&_th]:text-navy dark:[&_th]:text-white [&_td]:py-2 [&_td]:px-3 [&_td]:border-b [&_td]:border-slate/10 dark:[&_td]:border-white/10 [&_a]:text-teal dark:[&_a]:text-teal-bright [&_a]:underline"> TL;DR: we surveyed the public pricing of 15+ AI-visibility products and services in June 2026. One-time audits run $27–$4,500, and depth varies wildly — from 3 keywords to 100 prompts. Monitoring tools run $29–$654 per month, with a category average near $337. Enterprise platforms start at $399–499 per month behind a sales call. Agency GEO retainers run $2,000–8,000 per month. Depth-per-dollar varies by an order of magnitude. The tables below are the comparison we wish had existed. ## What do one-time audits cost? Offer type Price (June 2026) Scope Turnaround | | Automated scan (e.g. SeekON, TransCanada starter) | $27 | ~3 keywords, automated output | Instant | | Thin productized audit (e.g. MassMonopoly Starter) | $497 | 5–8 queries, 3 engines | ~5 days | | Mid productized audit (e.g. MassMonopoly Standard/Complete) | $1,497–2,497 | 15–40 queries, 4 engines | 5–7 days | | CitedMetrics audit (for comparison) | $950 | 50–150 prompts × 4 engines × 3 runs | 5 business days | | Agency-grade audit (e.g. Veza Digital) | $4,500 | 50–100 prompts, 6 platforms, 30-page report | ~6 weeks | | Custom-scoped consultancy audits | Quote (typ. $1,500–7,500) | Varies; industry guides put 10-prompt audits at $1,500–3,000 | 10–14+ days | ## What do monitoring subscriptions cost? Tool Entry price (June 2026) Notes | | Otterly.AI | from $29/mo | Self-serve | | Knowatoa | $59–199/mo | Self-serve, 7 engines | | Semrush AI toolkit | $99/mo per domain | Add-on | | AthenaHQ | $295/mo (self-serve) | Enterprise tier custom | | Ahrefs Brand Radar | ≈€358–654/mo | Within Ahrefs plans | | Peec AI | €85–425/mo | +€30–140/mo per extra engine | | Profound | $399–499/mo entry; $2,000–5,000+/mo enterprise | Sales-call gated | | Category average | ≈$337/mo | Hamster Garage tools benchmark | ## What do agency retainers cost? GEO and AEO implementation retainers cluster at $2,000–8,000 per month, usually with 3-month minimums. Most include reporting. Few include a rigorous multi-run diagnostic up front. That is why audit-first, then implement — in-house or via retainer — is the cost-rational order. ## How should you read this market? - Depth-per-dollar varies ~10×. $1,497 can buy 15–20 queries; $950 can buy 150 prompts × 3 runs. Always ask for prompt counts and run counts in writing. - Subscriptions price in your labor. Dashboards under ~$300/mo typically require you to design prompts and interpret output. Budget your hours alongside the fee. - Speed is rare at depth. Above ~$2,500, turnarounds stretch to weeks. Below $500, depth collapses to a handful of queries. Methodology note: we collected prices from public pricing pages and published reviews in June 2026. Every source is listed in the citations. Where pricing sits behind a sales call, we used published third-party reporting. This market shifts fast — treat the survey as a dated snapshot, exactly as we'd tell you to treat any AI-visibility number. ## Frequently asked questions - What does an AI visibility audit cost in 2026? One-time audits range from $27 (automated scans of a few keywords) to $4,500 (agency-grade, 50–100 prompts, ~6-week delivery). Credible human-analyzed audits cluster between $950 and $2,500. Monitoring subscriptions average ≈$337/month; agency GEO retainers run $2,000–8,000/month with 3-month minimums. - How much does an audit usually cost? For comparison, traditional SEO audits typically run $2,500–5,000 as projects, with agency versions reaching $10,000. AI-visibility audits are younger and cheaper at equal depth — but the spread is wider, so compare prompt counts and run counts, not price tags. - What is a reasonable audit fee? Price it per unit of evidence: prompts × engines × runs. At $950 for 50–150 prompts × 4 engines × 3 runs, you pay well under a dollar per scored answer. A $1,497 audit covering 15–20 single-run queries costs many times more per data point. Reasonable means evidence-dense, not cheap. - Audit or monitoring subscription — which first? Audit first. A subscription tells you numbers move; a diagnostic tells you why and what to fix. Most monitoring tools also require you to configure prompts and interpret results yourself — pricing in your hours on top of the fee. ## Keep reading Research What is an AI search visibility audit? Research CitedMetrics audit methodology Guides GEO vs SEO: what actually changed? Written by David King · Founder I built CitedMetrics to create the AI-visibility measurement I wished existed for my own 11 content sites. Before that I managed aerospace manufacturing at Rolls-Royce — overseeing the build and assembly of complete jet engine sections for Airbus and Boeing aircraft — where no decision was made without extensive data analysis. I apply the same standard here: every finding in your audit is backed by real prompts, real engine responses, and scored evidence. --- # What is an AI search visibility audit? Source: https://citedmetrics.com/research/what-is-an-ai-search-visibility-audit/ - Home - / Research - / What is an AI search visibility audit? # What is an AI search visibility audit? Published 2026-06-10 · Updated 2026-06-10 · David King h2]:font-headline [&>h2]:text-2xl [&>h2]:font-bold [&>h2]:text-navy dark:[&>h2]:text-white [&>h2]:mt-10 [&>h2]:mb-3 [&>ul]:list-disc [&>ul]:pl-6 [&>ul]:space-y-2 [&_table]:w-full [&_table]:text-sm [&_table]:border-collapse [&_th]:text-left [&_th]:py-2 [&_th]:px-3 [&_th]:border-b-2 [&_th]:border-slate/20 dark:[&_th]:border-white/20 [&_th]:font-bold [&_th]:text-navy dark:[&_th]:text-white [&_td]:py-2 [&_td]:px-3 [&_td]:border-b [&_td]:border-slate/10 dark:[&_td]:border-white/10 [&_a]:text-teal dark:[&_a]:text-teal-bright [&_a]:underline"> TL;DR: an AI search visibility audit measures what AI engines — ChatGPT, Claude, Gemini, Perplexity — tell buyers about a brand. It runs real buying-intent prompts repeatedly, scores every answer (mentioned, cited, recommended, or invisible), benchmarks competitors, verifies what the engines claim, and ends in a prioritized fix roadmap. It is a diagnosis with evidence — not a score, and not a dashboard. ## What does it actually measure? - Mention rate — how often the brand appears in answers at all. - Recommendation rate — how often engines endorse it as a pick. - Citation rate — how often the brand's own site is used as a source. - Share of voice — the brand's slice of all brand mentions in its category. - Accuracy — whether what engines claim (prices, products, capabilities) is true. ## What separates a real audit from a scan? Repetition and breadth. Engines answer differently run to run, so a real audit runs every prompt several times and reports the variance. It covers enough prompts to map a funnel — comparison, validation, problem, alternatives — not five queries. It tests multiple engines, because they disagree. And it benchmarks competitors, because a score without share-of-voice context means nothing. A $27 scan does none of this; it prints a number. ## What arrives at the end? A report: the scorecard, share-of-voice charts, prompt-by-prompt results, flagged claims with raw evidence, technical root causes (crawler access, schema, entity consistency), and a sequenced roadmap. Ours also ships the raw data export and 90 days of an interactive dashboard — see a full sample built from real engine runs. ## What does one cost? Our June 2026 market survey maps the range: $27 automated scans, $497–2,497 thin audits covering 5–40 prompts, $950 for the CitedMetrics 50–150-prompt 4-engine audit, $4,500+ agency engagements, and monitoring tools averaging ≈$337/month forever. The honest comparison metric is depth per dollar: prompt count × engines × runs. ## Who actually needs one? Any company whose buyers research before buying — which in B2B is measurably most of them: 84% of B2B SaaS CMOs now use LLMs in vendor discovery (Wynter, 2026; the numbers live in our buyer-behavior brief ). If the shortlist forms inside AI answers and you've never measured your presence there, the audit is the instrument. The free snapshot is the smallest honest version of it. ## Frequently asked questions - How is AI visibility calculated? As rates over repeated runs: mention rate (runs naming the brand ÷ total runs), citation rate (runs sourcing the brand's site), recommendation rate (runs endorsing it), and share of voice (the brand's mentions ÷ all brand mentions in the category). Single runs are excluded by design — engines vary, so only repeated runs measure. - How much does an AI visibility audit cost? In June 2026: $27 automated scans (a few keywords, no analysis), $497–2,497 thin productized audits (5–40 prompts), $950 for our 50–150-prompt 4-engine audit, and $4,500+ for agency-grade engagements with multi-week turnarounds. Depth-per-dollar varies by an order of magnitude — always ask for prompt and run counts in writing. - Is an audit better than a monitoring tool? Different jobs. A monitoring dashboard (≈$337/month category average) tracks numbers over time but leaves prompt design and interpretation to you. An audit is a diagnosis: what is true today, why, and what to fix first. The cost-rational order is audit first, then monitor the fixes. ## Keep reading Research AI visibility audit & tool pricing — June 2026 market survey Research CitedMetrics audit methodology Guides How to check what AI says about your brand Written by David King · Founder I built CitedMetrics to create the AI-visibility measurement I wished existed for my own 11 content sites. Before that I managed aerospace manufacturing at Rolls-Royce — overseeing the build and assembly of complete jet engine sections for Airbus and Boeing aircraft — where no decision was made without extensive data analysis. I apply the same standard here: every finding in your audit is backed by real prompts, real engine responses, and scored evidence. --- # Refund Policy Source: https://citedmetrics.com/refund-policy/ - Home - / Refund Policy # Refund Policy Last updated: 10 June 2026 h2]:font-headline [&>h2]:text-2xl [&>h2]:font-bold [&>h2]:text-navy dark:[&>h2]:text-white [&>h2]:mt-10 [&>h2]:mb-3 [&>ul]:list-disc [&>ul]:pl-6 [&>ul]:space-y-2 [&_a]:text-teal dark:[&_a]:text-teal-bright [&_a]:underline"> We sell a one-time AI Visibility Audit for $950 USD. This policy describes exactly when you get your money back. It applies to every purchase made through this website. ## Can I cancel before the work starts? Yes — full refund. If you cancel before we begin running your prompts (work begins after you return the intake form), email support@citedmetrics.com from the address used at checkout and we refund 100% of your payment, no questions asked. ## What if my report is late? Your report is due by email within 5 business days of receiving your completed intake form (see our Delivery Policy ). If we miss that window and you no longer want the audit, you can request a full refund — whether or not the work is in progress. ## What if the report doesn't teach me anything? (Keep-the-report guarantee) Read the full report. If it doesn't show you something about your AI visibility you didn't already know, reply to the delivery email within 14 days and we refund the full $950 — and you keep the report. We can offer this because every finding is backed by raw engine responses; the work speaks for itself. ## How are refunds processed? - Refunds go back to the original payment method via Stripe. - We confirm by email within 2 business days of your request; funds typically appear in 5–10 business days, depending on your bank. - No partial refunds, no credit-only refunds — refunds under this policy are always the full purchase amount. ## How do I request a refund? Email support@citedmetrics.com (or simply reply to your order or delivery email) with the email address used at checkout. No forms, no calls. --- # Delivery Policy Source: https://citedmetrics.com/delivery/ - Home - / Delivery Policy # Delivery Policy Last updated: 10 June 2026 h2]:font-headline [&>h2]:text-2xl [&>h2]:font-bold [&>h2]:text-navy dark:[&>h2]:text-white [&>h2]:mt-10 [&>h2]:mb-3 [&>ul]:list-disc [&>ul]:pl-6 [&>ul]:space-y-2 [&_a]:text-teal dark:[&_a]:text-teal-bright [&_a]:underline"> The AI Visibility Audit is a digital service. Nothing ships physically — everything arrives by email. ## What do I receive? - A 20+ page PDF report: visibility baseline, competitor share-of-voice, sentiment and hallucination findings with evidence, technical root-cause analysis, and a prioritized 60–90-day fix roadmap. - The raw scored data export of every engine response, so your team can verify the findings. ## When do I receive it? - Day 0: immediately after payment you receive an order confirmation and a 5-minute intake form (your domain, your closest competitors, anything you want probed). - Days 1–4: we run 50–150 buying-intent prompts across ChatGPT, Claude, Gemini and Perplexity — three runs each — and score every answer. - Day 5: your report and raw data arrive in the inbox used at checkout, within 5 business days of receiving your completed intake form. If anything threatens that window (an engine outage, an unusually large prompt set), we tell you by email before the deadline — and a missed window qualifies you for a full refund under the Refund Policy . ## What format does everything arrive in? The report is a PDF — readable on anything, forwardable to anyone. The raw data export arrives alongside it as spreadsheet-friendly files, one row per prompt per engine per run, so your analyst can filter, chart, and re-check our scoring. No portal, no login, no software to install. ## What about the free snapshot? The free 10-prompt snapshot is delivered to your inbox within 2 business days of your request. It is a 1-page summary, not the full audit. ## Questions? Email support@citedmetrics.com — we answer within one business day. --- # Privacy Policy Source: https://citedmetrics.com/privacy/ - Home - / Privacy Policy # Privacy Policy Last updated: 10 June 2026 h2]:font-headline [&>h2]:text-2xl [&>h2]:font-bold [&>h2]:text-navy dark:[&>h2]:text-white [&>h2]:mt-10 [&>h2]:mb-3 [&>ul]:list-disc [&>ul]:pl-6 [&>ul]:space-y-2 [&_a]:text-teal dark:[&_a]:text-teal-bright [&_a]:underline"> CitedMetrics ("we") sells AI visibility audits. We collect the minimum data needed to deliver them, and we don't sell or rent any of it. This page lists everything we collect and why. ## What do we collect? - Free snapshot form: your company domain and work email — used only to generate and deliver your snapshot, and to follow up about it once. - Audit purchases: your name, email, company domain and the intake details you provide (competitors, product category) — used only to produce and deliver your audit. - Payments: handled entirely by Stripe. We never see or store your card details. Stripe's privacy policy is at stripe.com/privacy . - Analytics: none at present. No advertising trackers, no cross-site cookies. If we add analytics it will be a cookieless, privacy-respecting service and this policy will be updated first. ## How is audit data used? - Your domain and category terms are included in prompts sent to third-party AI engines (OpenAI, Anthropic, Google, Perplexity) via measurement APIs to produce your results. That is the product. - Your report, prompts and results are never shared, resold, or used in our marketing without written permission. - We keep audit data while you remain a client and for up to 24 months after, so re-runs can be compared against your baseline; ask and we delete it sooner. ## Who do we share data with? - Stripe — payment processing. - Our transactional email provider — to send your snapshot, report and receipts. - AI measurement APIs — your domain/category terms appear inside prompts, as described above. - Nobody else. No data brokers, no advertisers. ## What are your rights? Email support@citedmetrics.com to access, correct, export, or delete any personal data we hold about you. We action requests within 30 days. If you are in the UK/EEA, you also have the right to complain to your supervisory authority. ## Who is the data controller? CitedMetrics, Business address available on request. Contact: support@citedmetrics.com . --- # Terms of Service Source: https://citedmetrics.com/terms/ - Home - / Terms of Service # Terms of Service Last updated: 10 June 2026 h2]:font-headline [&>h2]:text-2xl [&>h2]:font-bold [&>h2]:text-navy dark:[&>h2]:text-white [&>h2]:mt-10 [&>h2]:mb-3 [&>ul]:list-disc [&>ul]:pl-6 [&>ul]:space-y-2 [&_a]:text-teal dark:[&_a]:text-teal-bright [&_a]:underline"> These terms govern your purchase of an AI Visibility Audit from CitedMetrics (Business address available on request). Buying the audit through our Stripe checkout means you accept them. ## What is the service? A one-time diagnostic: 50–150 buying-intent prompts run across ChatGPT, Claude, Gemini and Perplexity (three runs each), scored for mentions, citations and recommendations; competitor share-of-voice; sentiment and hallucination findings; technical root-cause analysis; and a prioritized 60–90-day fix roadmap. Deliverables and timing are described in the Delivery Policy . ## What does it cost? $950 USD, one-time, paid in advance via Stripe. This is a founding rate; we may raise the listed price for future purchases (it will never change after you've paid). Prices exclude any taxes your jurisdiction applies to digital services. ## What do we need from you? A completed intake form with accurate information about your domain, category and competitors. The 5-business-day delivery clock starts when we receive it. If the intake is not returned within 30 days of purchase, we'll check in twice and then refund you in full rather than hold the payment indefinitely. ## What can you do with the report? The report and raw data are yours: use them internally, share them with your agency, quote them in board decks. You may not resell the report as a standalone product or pass it off as your own audit service. Our methodology, prompt libraries and scoring rubrics remain ours. ## What are the limits of the findings? - AI engines are third-party services that change continuously. Your report is a dated, multi-run measurement — a baseline, not a permanent state. - Engine answers quoted in the report are third-party output, reproduced as evidence. They are not our claims about you or your competitors. - The roadmap is professional advice based on documented citation factors. We can't guarantee specific visibility outcomes — anyone who does is selling vibes, and we recommend reading our research on how to judge such claims. ## Refunds and cancellation Covered in full in the Refund Policy : cancel any time before work begins, full refund if delivery is late, and a 14-day keep-the-report guarantee after delivery. ## Liability Our total liability for any claim arising from the service is capped at the amount you paid. We are not liable for indirect or consequential losses (lost profits, lost rankings, decisions made on third-party engine output). Nothing here limits liability that cannot lawfully be limited. ## Questions or disputes Email support@citedmetrics.com first — most things resolve in one reply. Formal disputes are governed by the laws applicable to the registered business address provided on your order correspondence.