# Cited By AI® # AI Search Engine Optimisation (ASEO) Consultancy (UK) # https://citedbyai.info # Updated: 2026-06-19 > Cited By AI® is a UK-based ASEO consultancy that helps businesses get cited by ChatGPT, Perplexity, Claude, Gemini, and Microsoft Copilot. We produce a 28-module AI visibility audit covering Share of Voice, funnel-stage SOV, Citation Probability Score per page, hallucination detection, competitive win rate, publish-ready AEO content, and Bing IndexNow submission. As of Q1 2026, the framework covers 28 modules producing a 35-section report. Our registered trademarks include Cited By AI® and Citation Probability Score (CPS). ## Core Pages - [Home](https://citedbyai.info/): Main site: overview of ASEO services, audit modules, pricing (Solo audits from £49 one-off or £99-149/month with no minimum commitment), and the free automated CPS Lite Score - [Why Us / How We Compare](https://citedbyai.info/why-us): Feature-by-feature comparison of our approach vs other AI visibility products ## Services - AI Visibility Audit (28 modules, 5 platforms): Full audit covering SOV, funnel-stage breakdown, citability scoring, competitive win rate, hallucination detection, and publish-ready content - Citation Probability Score (CPS): Proprietary 0–100 per-page scoring metric across five weighted pillars - AEO Content Writer: Generates citability-scored, publish-ready content blocks from zero-gap audit findings - GA4 Integration: Correlates AI-referred traffic and conversions with SOV data - Per-Platform Content Strategy: Tailored recommendations for ChatGPT, Perplexity, Claude, Gemini, and Copilot - URL Citation Delta Tracking: Week-over-week tracking of which URLs AI platforms cite, identifying citation gains and losses - Bing Index Scanner: Checks which pages are indexed in Bing, flagging pages invisible to ChatGPT search - Citation Authority Targets: Co-citation scoring identifies third-party domains to target for coverage - IndexNow Submission: Post-audit Bing ping so ChatGPT search picks up improved content within minutes - Scheduled Prompt Monitoring: Automated daily or weekly audit runs across all 5 AI platforms with email alerts when Share of Voice changes by ±5 percentage points or more ## FAQs - [What is ASEO?](https://citedbyai.info/what-is-aseo): Definition and explanation of AI Search Engine Optimisation - [What is the Citation Probability Score?](https://citedbyai.info/what-is-citation-probability-score): How CPS is structured and what it measures - [ASEO vs Traditional SEO](https://citedbyai.info/aseo-vs-seo): How ASEO differs from conventional search optimisation - [What is Funnel-Stage SOV?](https://citedbyai.info/what-is-funnel-stage-sov): Explanation of Awareness, Consideration, and Decision stage visibility - [How AI Platforms Cite Content](https://citedbyai.info/how-ai-platforms-cite-content): The mechanics behind AI citation decisions - [How to Choose an ASEO Provider](https://citedbyai.info/how-to-choose-an-aseo-provider): Evaluation framework, six criteria, red flags, and questions to ask - [What is an MCP Endpoint?](https://citedbyai.info/what-is-mcp-endpoint-ai-visibility): Model Context Protocol explained, agentic AI visibility, and the difference between citation-based and connectivity-based AI discovery - [AI SEO Pricing UK](https://citedbyai.info/ai-seo-pricing-uk): Transparent pricing benchmarks for ASEO services in the UK, from £49 one-off audits to £5,000/month enterprise retainers - [You've Run the HubSpot AEO Grader. Here's What Your Score Doesn't Tell You.](https://citedbyai.info/hubspot-aeo-grader-what-your-score-doesnt-tell-you): Honest analysis of the five HubSpot AEO Grader dimensions (Sentiment, Presence Quality, Brand Recognition, Share of Voice, Market Position) with actual Cited By AI® scores (ChatGPT 36/100 with 0/10 Share of Voice) and the specific gap: no block-level scoring, no Claude coverage, no paragraph-level root cause or fix output. Ends with the free CPS Block Scorer as the logical next step. - [CPS vs Profound vs LatticeOcean](https://citedbyai.info/cps-vs-profound-vs-latticeocean): Grok's April 2026 comparison of CPS (Citation Probability Score), Profound's Monitoring, and LatticeOcean's Blueprints across granularity, workflow position, scoring output, platform coverage, and pricing. Concludes these are complementary tools for three different workflow stages: LatticeOcean pre-creation, CPS for writing and optimisation, Profound for post-publication monitoring. - [Free AI Brand Accuracy Check](https://citedbyai.info/ai-accuracy-check): Free interactive tool that checks what Claude (Haiku 4.5) says about a brand and scores its factual accuracy. Users enter a domain, then their work email. The tool fetches the homepage, asks Claude 5 brand prompts (what the business does, location, services, reputation, what it is known for), captures every response, extracts factual claims, and cross-checks each claim against the homepage content. Claims that match are marked verifiable; claims with no homepage anchor are flagged as unverified. Results show an overall accuracy percentage, a per-prompt breakdown with claim counts, and a sample unverified claim for each prompt that surfaces one. One free check per domain, ever. Gated by Turnstile, email validation, domain-email matching, and IP rate limiting. Upgrades to the paid AI Brand Accuracy Report (5 platforms, 25 prompts, judge-scored against Verified Brand Facts) via /ai-brand-accuracy. - [AI Brand Accuracy: Stop AI Misrepresenting Your Brand](https://citedbyai.info/ai-brand-accuracy): Service and methodology page for the AI Brand Accuracy offering. Defines AI Brand Accuracy as the measurement of whether what AI platforms (ChatGPT, Perplexity, Claude, Gemini, Copilot) say about a brand is factually correct, distinguishing it from Citation Probability Score (CPS) which measures citation likelihood. Includes a six-row comparison table of CPS vs AI Brand Accuracy, a five-dimension accuracy framework (Services, Location, Contact, Identity, Credentials), and a section on Verified Brand Facts (the proprietary schema that serves as ground truth for every paid Accuracy Report). Three product tiers: Free Check (1 platform, 5 prompts, homepage verification), Full Accuracy Report (5 platforms, 25 prompts, judge-scored against signed Verified Brand Facts, docx + Excel + share URL), and Monthly Monitor (monthly re-audits with claim-level deltas and alerting). Contains a FAQPage schema with 6 structured questions covering the definition, difference from CPS, free check accuracy, paid report scope, monitoring cadence, and Verified Brand Facts schema. - [ASEO for Shopify, Webflow, Squarespace and Custom Sites](https://citedbyai.info/aseo-for-shopify-webflow-squarespace): Platform-neutral ASEO landing page targeting buyers who cannot use WordPress-only AEO tools. Explains the architectural reason WordPress AEO plugins (which require WordPress admin, database access, and the plugin ecosystem) cannot be installed on Shopify, Webflow, Squarespace, HubSpot CMS, or custom-built sites. Describes how the Cited By AI 28-module ASEO audit operates at the AI-response and content layer rather than the CMS layer, making it compatible with every platform. Includes platform-specific notes for Shopify (thin product page content, retrieval failure patterns), Webflow (clean HTML, JSON-LD injectable, content structure gaps), Squarespace (schema limitations, block length issues), HubSpot CMS (content depth vs block-level structure gap), and custom builds (code-ready schema snippets). Features a plugin vs. audit comparison table and a 6-column platform support grid. 5 FAQPage schema entries covering Shopify, Webflow, non-WordPress constraint, platform-neutral audit scope, and full platform list. - [What Your Automation Agency Isn't Telling You About AI Visibility](https://citedbyai.info/automation-agencies-ai-visibility): Buyer briefing for SaaS, e-commerce, HR tech, professional services and accounting brands that have hired automation agencies (CRM integration, workflow automation in n8n/Make.com/Zapier, AI agent deployment) and assumed that means they're "AI-ready." Argues that internal automation and external AI search visibility are two separate stacks. Automation agencies optimise inside the business (HubSpot, Salesforce, Pipedrive, Airtable, custom AI agents); ASEO consultancies optimise outside it (whether ChatGPT, Perplexity, Claude, Gemini and Copilot cite the brand when buyers ask category questions). Includes a side-by-side stack comparison, a seven-row scope table covering work location, measurement, improvement targets, beneficiaries, tools, output and failure modes, a three-step buyer self-check (buyer query test, brand description test, GA4 referral check), and the complementary fit framing. 5 FAQPage schema entries covering the difference between automation agencies and ASEO consultancies, whether internal AI agents improve external visibility, how to spot a visibility gap, whether to replace an automation agency, and what an AI visibility audit includes. - [Citability Score vs CPS: What Each Actually Measures](https://citedbyai.info/citability-score-vs-cps): Methodology comparison between AEO God Mode's Citability Score (a 10-signal A+ to F page-level grading system available as a Pro feature of the AEO God Mode WordPress plugin) and Cited By AI®'s CPS (Citation Probability Score), a 5-pillar block-level measurement framework that scores content chunks at the 134-167 word RAG retrieval level. Draws the distinction across nine dimensions: unit of measurement (page vs block), what each scores (10 publish-time signals vs 5 pillars + 23 audit modules), output format, platform coverage (WordPress-only vs platform-neutral), AI platforms queried (0 input-side vs 5 cross-platform), hallucination detection (absent vs included), funnel-stage SOV (absent vs included), revenue attribution (absent vs GA4 included), and delivery (self-serve plugin vs practitioner audit). Includes side-by-side score cards, a nine-row comparison table, a five-question buyer self-check, and the complementary-fit framing for WordPress operators who can run both. Positions CPS as the deeper measurement standard while treating Citability Score as a competent tool in its own category. 6 FAQPage schema entries covering the difference between the two scores, WordPress-only constraint, why block-level scoring matters versus page-level, hallucination detection scope, whether both can be used together, and what CPS stands for. - [What Actually Drives AI Citations, and What's Just Sold to You](https://citedbyai.info/what-actually-drives-ai-citations): Cornerstone synthesis resource bundling three independent 2026 studies that converge on the same conclusion: schema markup and llms.txt files are not the AI citation levers they are marketed as. Study 1 is the Ahrefs controlled schema study (Linehan & Guan, 11 May 2026): 1,885 pages adding JSON-LD against 4,000 matched controls, difference-in-differences analysis, no statistically significant citation uplift on Google AI Mode (+2.4%) or ChatGPT (+2.2%), small decline on AI Overviews (-4.6%). Study 2 is the Ahrefs llms.txt traffic study (Linehan & Guan, 15 June 2026): 137,000 domains analysed, 28% publish llms.txt, 97% of those files received zero requests in May 2026, and of the 3% read, AI retrieval bots (OAI-SearchBot, PerplexityBot) accounted for just 1.1% of requests while agentic infrastructure like Claude-Code was the largest AI consumer at 10.5%, indicating a developer-tooling use case rather than a citation one; roughly 0.6% of published llms.txt files ever receive any AI bot request. Includes the full 12-category bot breakdown table. Study 3 is Google Search Central's 15 May 2026 AI optimisation guide, whose Mythbusting section states neither structured data nor machine-readable files like llms.txt are required for generative AI search. The page argues these tactics persist because they are easy to measure, easy to generate, and correlate with citation in observational data (sites that implement them also invest in content quality), but controlled studies that isolate the variable find the correlation does not hold as causation. Positions CPS as the evidence-led alternative: block-level scoring across Content Structure, Fact Density, Answer Architecture, Self-Containment, and Freshness, mapping to the Princeton/KDD GEO benchmark top interventions (Cite Sources +40.6%, Quotation Addition +35.1%, Statistics Addition +32.9%). Links out to three deep-dive pages (schema-vs-content-signals-ai-citations, machine-readability-is-not-enough, google-ai-optimisation-guide-analysis) and the CPS Research Foundation rather than duplicating them. Includes an honest scope callout: schema retains crawlability and entity-recognition value, llms.txt is a legitimate Layer 3 agentic-readiness signal, neither is a citation lever on already-visible pages. 4 FAQPage schema entries and an Article citation array referencing both Ahrefs studies and Google Search Central. - [Schema vs Content Signals: What Actually Drives AI Citations](https://citedbyai.info/schema-vs-content-signals-ai-citations): Strategy briefing published 11 May 2026 in response to the Ahrefs controlled study (Linehan & Guan, 11 May 2026) that tracked 1,885 pages adding JSON-LD schema between August 2025 and March 2026 against 4,000 matched control pages and found no statistically significant citation uplift on Google AI Mode (+2.4%) or ChatGPT (+2.2%), with a small statistically significant decline on Google AI Overviews (-4.6%). Pairs the Ahrefs finding with the mechanistic searchVIU experiment showing ChatGPT, Claude, Perplexity, Gemini and Google AI Mode all ignore JSON-LD during direct retrieval. Separates the AI citation problem into two stacks: technical readiness (where schema lives: crawler access, llms.txt, knowledge graph signals) and citation decisions (what actually drives selection: block size in the 134-167 word RAG range, declarative opening sentences, fact density, self-containment, freshness markers). Reframes schema as a technical readiness tool, not a citation driver on already-visible pages. Presents the five CPS pillars as the content-level signals that research shows actually drive citation decisions. Includes three concrete strategy updates for buyers: stop treating schema as the primary citation lever, audit content at the block level, treat schema-first AI visibility tools as solving the technical readiness problem rather than the citation decision problem. Cites Ahrefs and searchVIU as structured Article schema citations. 6 FAQPage schema entries covering whether schema increases AI citations, why 53% of cited pages have schema if it doesn't drive citations, whether AI systems read schema during retrieval, whether to stop implementing schema, what actually drives citations, and what to update in an ASEO strategy after this study. - [AEO vs GEO vs ASEO: The Definitive Comparison for 2026](https://citedbyai.info/aeo-vs-geo-vs-aseo): Definitional comparison piece treating AEO (Answer Engine Optimisation), GEO (Generative Engine Optimisation), and ASEO (AI Search Engine Optimisation) as three terms describing overlapping disciplines at different scopes. Documents the formal origin of GEO in Aggarwal et al., "GEO: Generative Engine Optimization," KDD 2024 (Princeton, IIT Delhi, Georgia Tech, Allen Institute), arXiv preprint November 2023 (arXiv:2311.09735), DOI 10.1145/3637528.3671900, which tested six content modification strategies across 10,000 queries on 10 generative engines and found statistics addition improved visibility 41%, citing external sources improved visibility 115% for lower-ranked content, quotations 28%, and keyword stuffing performed 10% worse than baseline. Notes Wikipedia's explicit statement that no consensus definition distinguishing these terms had been established in peer-reviewed literature as of early 2026. Treats AEO as the oldest term (pre-LLM, featured snippets and voice assistants), GEO as the peer-reviewed academic term, and ASEO as the broadest operational scope covering crawler access, block-level citation scoring, hallucination detection, funnel-stage Share of Voice, competitive win rate, and GA4 revenue attribution. Includes a 13-row comparison table across full name, first formal use, origin context, primary unit, core measurement, platforms, schema emphasis, hallucination detection, revenue attribution, funnel-stage SOV, tools using each term, and operational scope. Includes three deep-dive sections (one per term) with strengths and limits, and a five-question buyer framework based on CMS, measurement vs optimisation scope, platform count, hallucination needs, and revenue attribution. Positions ASEO as the most operationally complete framework while treating AEO and GEO as competent terms used by competent practitioners in adjacent specialisations. Schema graph includes ScholarlyArticle citation to the Aggarwal et al. KDD paper. 6 FAQPage schema entries covering the difference between AEO, GEO and ASEO, who coined GEO, whether AEO and GEO are the same, which framework a business should use, why Cited By AI uses ASEO, and whether the three frameworks replace traditional SEO. - [Platform-Specific Citation Guidance: How Each of the Five AI Platforms Actually Works](https://citedbyai.info/platform-specific-citation-guidance): Definitive per-platform breakdown of how Google AI Overviews and AI Mode, Microsoft Copilot, ChatGPT Search, Perplexity, and Claude each handle retrieval, query decomposition, and citation. Google's 15 May 2026 AI optimisation guide explicitly confirms RAG (retrieval-augmented generation) and query fan-out as the mechanisms behind its generative AI features, layered on its core Search ranking systems. Microsoft Copilot uses Bing's index (10 billion+ pages) with GPT-4 synthesis and now exposes grounding queries in Bing Webmaster Tools' AI Performance Report dashboard (launched by Krishna Madhavan, Meenaz Merchant, Fabrice Canel, and Saral Nigam). ChatGPT historically used Bing as its primary search index from October 2023 but is transitioning toward proprietary retrieval as of March 2026, with OAI-SearchBot and ChatGPT-User as active retrieval bots (separate from GPTBot, which is training-only). Perplexity owns its own search index of over 200 billion URLs built on Vespa, with tens of thousands of index updates per second, chunk-level scoring, and the proprietary Sonar model based on Llama. Claude defaults to its training corpus on claude.ai with Brave Search as the web search backend when invoked, confirmed by Anthropic's subprocessor documentation and Profound's 2025 finding of 86.7% citation overlap between Claude responses and Brave's top results; ConvertMate research found 68% of Claude's factual citations come from traditional structured databases (Wikipedia, academic, government, business directories) with 70% verified across multiple authoritative sources. Includes a 10-row comparison table across retrieval index, default search behaviour, query decomposition, unit of retrieval, schema weight, llms.txt weight, freshness weight, entity authority weight, active retrieval bot, and citation behaviour. Three strategy implications: you can't use Google's playbook on Claude, you can't use Claude's playbook on Perplexity, and one averaged score across five different mechanisms is a misleading diagnostic. 6 FAQPage entries covering whether Google's guide applies to other platforms, why CBA's audit covers all five, which platform is hardest to get cited by, which is most freshness-sensitive, whether Bing visibility still drives ChatGPT citations in 2026, and what query fan-out is and which platforms use it. Article schema citation array references Google Search Central, Perplexity Research, Anthropic, and Microsoft as primary sources. - [Google's AI Optimisation Guide: What It Says, What It Doesn't, and What It Means for ASEO](https://citedbyai.info/google-ai-optimisation-guide-analysis): Analytical response piece to Google Search Central's official 15 May 2026 guide titled "Optimising your website for generative AI features on Google Search." Treats the document as the most authoritative platform statement of the year so far, deserving careful reading rather than trade-press weaponisation. Confirms two retrieval mechanisms Google explicitly named: retrieval-augmented generation (RAG) and query fan-out, with Google's own lawn-care example showing fan-out queries 'best herbicides for lawns', 'remove weeds without chemicals', and 'how to prevent weeds in lawn' generated from the single user query 'how to fix a lawn that's full of weeds'. Observed AI Mode behaviour shows up to 16 fan-out queries per single user search. Covers Google's four mythbusts (schema markup, llms.txt and similar AI text files, chunking content into tiny pieces, rewriting content specifically for AI systems) with Google's exact quotes and CBA's analytical position on each. Argues five claims: (1) Google's guide applies to Google only and doesn't speak to ChatGPT, Perplexity, Claude, or Microsoft Copilot, which use different retrieval architectures; (2) RAG and query fan-out validate the content-first methodology behind CPS, with the Self-Containment and Answer Architecture pillars now formally backed by first-party platform documentation; (3) the schema and llms.txt mythbusting is consistent with controlled research (Ahrefs 11 May 2026 schema study of 1,885 pages, searchVIU mechanistic retrieval experiment, Wood's 50-protocol CryptoContent.dev audit DOI 10.5281/zenodo.19253709, the Ahrefs 15 June 2026 llms.txt traffic study of 137,000 domains finding 97% of files are never read, and Google's own statement, five independent sources from different methodologies pointing the same direction); (4) Google's 'still SEO' framing is methodologically correct but the visibility outcomes are different: only 53% of domains cited in AI Mode match the top 10 organic search results, only 35% of URLs match per E2M Solutions analysis, and Pew Research's controlled 68,000-query study found a 46.7% relative decline in clicks when AI Overviews appear; pages cited inside AI Overviews see a ~35% CTR uplift; (5) CBA's block-level CPS scoring is descriptive of how RAG retrieval mechanically operates rather than prescriptive guidance to fragment content, addressing Google's chunking mythbust head-on with the honest framing that the action items are paragraph rewrites for clarity in place, not deliberate splitting. Three things the guide doesn't cover: the other four platforms, hallucination detection, and funnel-stage Share of Voice. Sales narrative implication: Google has now officially confirmed that a single user query generates multiple concurrent sub-queries; brands appearing only against the primary query are invisible to most of the fan-out, validating CBA's Zero-Gap Topic Matrix module with first-party documentation rather than competing-vendor research. 6 FAQPage entries with detailed answers. Article schema citation array references Google Search Central guide, Google Search Central blog announcement, Ahrefs Linehan & Guan ScholarlyArticle, and Wood CryptoContent CreativeWork with DOI. - [How to Build an ASEO Strategy in 2026: Beyond AEO, Beyond GEO](https://citedbyai.info/how-to-build-an-aseo-strategy): The definitive 9-phase ASEO strategy framework, distinguishing ASEO (AI Search Engine Optimisation) from AEO (Answer Engine Optimisation) and GEO (Generative Engine Optimisation). Phases: 1) Platform audit and Share of Voice baseline across 5 platforms, 2) Platform-specific citation behaviour: detailed table comparing ChatGPT (Bing retrieval, FAQ schema), Perplexity (real-time, freshness-sensitive), Gemini (Google index, E-E-A-T), Claude (training knowledge base, fact density), and Copilot (Bing commercial intent), 3) Funnel-stage SOV mapping (Awareness/Consideration/Decision table with query patterns, what AI cites, and intervention type), 4) CPS block-level audit covering 5 pillars with weights and grade tier table, 5) Hallucination risk assessment with structured data interventions, 6) Zero-Gap Topic Matrix content planning, 7) CPS-scored content production, 8) Machine-readable infrastructure (llms.txt, Schema.org, sitemap, AI crawler access), 9) Monthly measurement loop with GA4 revenue attribution. Includes 5 FAQs in schema. Explicitly positions itself as the category upgrade to generic AEO strategy guides. - [What Comes After Your AI Visibility Score?](https://citedbyai.info/what-comes-after-your-ai-visibility-score): Landing page targeting users of Traqer, Peec AI, and Profound who have a visibility score but no diagnosis of what is causing it. Argues that monitoring tools (Traqer, Peec, Profound) measure share of voice at brand level while AI retrieval operates at paragraph level, and that the CPS audit bridges the gap. Includes a 12-row comparison table of what monitoring tools measure versus what a CPS audit diagnoses, covering block-level paragraph scoring, retrieval failure diagnosis, funnel-stage SOV, hallucination detection, GA4 revenue attribution, prioritised rewrite list, and Zero-Gap Topic Matrix. Core argument: monitoring tells you the score went down; a CPS audit tells you which sentence caused it and what to rewrite. Lists the 28-module audit outputs and explains how monitoring and auditing are complementary rather than competing. - [The Four Types of AI Brand Appearance, and Which Ones Actually Matter for Revenue](https://citedbyai.info/the-four-types-of-ai-brand-appearance): Extends Traqer's mention-versus-citation distinction with two additional types: hallucinations (factually incorrect AI statements about the brand) and sentiment frames (how AI positions the brand relative to competitors). Defines all four types: brand mentions (recommended in AI output, tracked by Traqer/Peec/Profound), citations (URL-sourced references, highest revenue relevance, generated by CPS-structured content), hallucinations (factually incorrect claims, highest risk, tracked by CBA AI Accuracy Audit), and sentiment frames (relative competitive positioning in AI responses, tracked by CBA Competitive Win Rate module). Includes a four-row table showing what moves each type, the measurement tool for each, and the typical timescale for improvement. Core argument: a share-of-voice score that counts hallucinations as positive appearances is actively misleading. All four types must be tracked separately. - [Does AI Content Writing Get You Cited?](https://citedbyai.info/does-ai-content-writing-get-you-cited): Answers the evaluative question buyers are asking before choosing an AI writing tool. Argues that AI content generation and AI citation verification are two separate operations, and that most tools only do the first. Explains the five block-level retrieval signals that determine whether a specific paragraph gets cited by ChatGPT, Perplexity, or Gemini: Content Structure (134-167 word RAG chunk, declarative opening), Fact Density (verifiable claims per 100 words), Answer Structure (declarative pattern matching query intent), Self-Containment (no cross-references or dangling pronouns), and Freshness Signals (date markers and recency language). Includes a side-by-side example of a non-citable vs citable paragraph on the same topic, the CPS grade tier table, and the distinction between content that reaches Grade B (65+/100) across all five pillars versus content that assumes citability without measurement. - [CPS Research Foundation: The Evidence Behind AI Citation Visibility](https://citedbyai.info/cps-research-foundation): The public research foundation for the Citation Probability Score (CPS) framework. Documents the evidence base for each of the five CPS pillars: Content Structure (40% visibility uplift in structured content per GEO KDD 2024), Fact Density (statistics boost AI citation by ~41% per GEO study), Answer Architecture (early-page content accounts for a disproportionate share of citations; as of May 2026 the pillar cites Google Search Central's own AI optimisation guide as first-party platform confirmation of query fan-out, with Google's example showing one query 'how to fix a lawn that's full of weeds' triggering concurrent sub-queries 'best herbicides for lawns', 'remove weeds without chemicals', 'how to prevent weeds in lawn'), Self-Containment (RAG chunk-level retrieval evaluation; the pillar now cites Google Search Central's first-party confirmation of retrieval-augmented generation as the grounding mechanism behind Google's generative AI features), and Freshness (Ahrefs 17M citation study: AI-cited content is significantly newer than ranked content). Includes cross-pillar authority evidence, a comparison of CPS vs traditional SEO vs GEO frameworks, and a published research caveat noting that some findings are correlational rather than causal. As of May 2026, the page presents disconfirming evidence from five independent sources alongside confirming evidence: (1) a controlled Ahrefs study (Linehan & Guan, 11 May 2026) of 1,885 pages adding JSON-LD schema against 4,000 matched control pages found no statistically significant citation uplift on Google AI Mode (+2.4%) or ChatGPT (+2.2%), with a small statistically significant decline on Google AI Overviews (-4.6%) on already-cited pages; (2) a searchVIU experiment found that ChatGPT, Claude, Perplexity, Gemini and AI Mode all ignore JSON-LD during direct retrieval, extracting only visible HTML; (3) an independent 50-protocol cross-sectional crypto audit by Wood (CryptoContent.dev, May 2026, DOI 10.5281/zenodo.19253709) corroborated the Ahrefs finding from a different vertical and methodology, showing that protocols with the highest schema quality scores were not the most cited (Pendle, the only DeFi protocol with JSON-LD, scored zero across all three platforms; Starknet with joint-highest schema quality ranked 15th overall; Aave with no schema ranked 2nd) and that official protocol pages accounted for just 1% of Perplexity citations and 4% of Google AI Overview citations across 1,016 records; (4) Google Search Central's official AI optimisation guide (15 May 2026) includes a 'Mythbusting' section explicitly stating that schema markup, llms.txt and similar AI text files, chunking content into tiny pieces, and rewriting for AI are all unnecessary for visibility in Google's generative AI features; the guidance covers Google's systems only (AI Overviews and AI Mode), and CBA's framework explicitly notes that ChatGPT, Perplexity, Claude and Microsoft Copilot use different retrieval and citation architectures. As of late May 2026, the page also adds an independent third-party synthesis card citing Perea Research's GEO/AEO 2026 paper (Dante Perea, 6 May 2026, CC BY 4.0 licensed), which draws on 100+ primary sources including the Princeton/KDD GEO benchmarks and the 5W AI Platform Citation Source Index. Perea's piece independently validates two specific CBA design choices: (1) the five-engines-five-citation-logics finding validates CBA's per-platform audit scope, with ZipTie cross-platform data showing only 11% citation overlap between ChatGPT and Perplexity and 71% of cited sources appearing on a single platform; (2) the seven-factor citation hierarchy validates the CPS five-pillar structure, with the top three Princeton/KDD interventions (Cite Sources +40.6%, Quotation Addition +35.1%, Statistics Addition +32.9%) mapping directly onto Content Structure, Fact Density, Answer Structure, Self-Containment, and Freshness. The Wood study also appears in Cross-Pillar Evidence as positive supporting evidence that AI citation is shaped by the distributed third-party content record surrounding an entity rather than what the entity publishes about itself (YouTube was the top cited domain in Google AIO with 111 citations, Reddit 94, CoinGecko led Perplexity citations). CBA frames schema and llms.txt as crawlability and entity-recognition tools in its technical readiness layer rather than as direct citation drivers, with explicit platform-specific guidance distinguishing Google AI surfaces from the other four platforms tracked. Reviewed quarterly. - [Cited By AI® vs LightSite AI: Consultancy vs. Platform](https://citedbyai.info/cited-by-ai-vs-lightsite): Direct comparison of Cited By AI® (UK ASEO consultancy, 27-module diagnostic audit, block-level CPS scoring, hallucination detection, GA4 revenue attribution, human expert delivery) against LightSite AI (Tel Aviv GEO infrastructure platform, $129-$299/yr, machine-readable layers, AI sitemaps, JSON-LD, agentic monitoring and content execution). Frames the comparison as consultancy vs. platform operating at different layers of the AI search stack: CBA at the selection layer (which paragraphs get cited and why), LightSite at the eligibility layer (can AI systems parse the site). Includes a 13-row capability comparison table. LightSite pricing verified against public pages April 2026. - [ASEO vs GEO: What's the Difference?](https://citedbyai.info/aseo-vs-geo): Definitive explainer distinguishing ASEO (AI Search Engine Optimisation, a registered trademark of Cited By AI® in the UK) from GEO (Generative Engine Optimisation, coined in the Princeton/IIT arXiv paper 2311.09735, accepted KDD 2024). GEO is a content visibility framework addressing appearance rate in generative engine responses through content modification techniques. ASEO is a complete commercial discipline covering Share of Voice measurement, block-level CPS scoring, funnel-stage SOV, hallucination detection, competitive win rate, and GA4 revenue attribution. Includes a 10-row capability comparison table, source citations to the original GEO paper, and a buyer evaluation framework. LightSite AI named as primary commercial GEO platform. - [AEO Content Writer vs Generic AI Writers](https://citedbyai.info/aeo-content-writer-vs-generic-ai-writers): Explains why the Cited By AI® AEO Content Writer is ASEO-native rather than a generic AI writing tool. Generic AI writers generate content and hope AI platforms cite it; CBA's AEO Content Writer scores every content block against the five-pillar CPS framework before delivery, only releasing blocks at Grade B (65+/100) or above. Covers the mechanism, the five CPS pillars measured per block, a side-by-side comparison with generic AI writers, and how the tool fits within the full 27-module audit pipeline. References RankBuilder as the archetypal generic AI writer. - [Machine-Readability Is the Floor, Not the Ceiling](https://citedbyai.info/machine-readability-is-not-enough): Explains why machine-readability (schema, llms.txt, entity signals, crawler access) is Layer 1 eligibility: necessary but not sufficient for AI citation. Layer 2 is block-level citability, measured by CPS, which determines whether AI retrieval systems select your specific paragraphs over a competitor's. The five CPS pillars are Content Structure, Fact Density, Answer Structure, Self-Containment, and Freshness Signals. Now also covers the Lighthouse 13.3 distinction: Google Search Central says llms.txt isn't required for AI citation, while Google Lighthouse's new Agentic Browsing category audits for it as a Layer 3 readiness signal for browser automation agents. Three layers, three buyer questions: can the bots read my site, will AI cite my brand, can browser agents act on my site. - [BISCUIT vs CPS Framework Comparison](https://citedbyai.info/biscuit-vs-cps-framework-comparison): Honest side-by-side comparison of Knowatoa's BISCUIT framework and Cited By AI®'s Citation Probability Score. Covers what each measures, their unit of analysis (brand/page vs 134–167 word block), where they overlap, and where CPS goes deeper with paragraph-level diagnosis and specific rewrites. - [What Is AI Query Fan-Out?](https://citedbyai.info/what-is-query-fan-out): How AI platforms expand one search query into 20 related prompts across Awareness, Consideration, and Decision funnel stages, and why most brands are visible on some and invisible on most. Explains the citation gap and how the CPS audit closes it. - [How to Prove Your ASEO Work is Driving Revenue](https://citedbyai.info/how-to-prove-aseo-drives-revenue): Why AI-referred traffic converts at 14.2% vs 2.8% for organic, how the GA4 revenue loop works inside the Cited By AI® audit, and why no other ASEO platform connects citation improvement to actual revenue - [The GA4 AI Assistant Channel: What It Shows and What It Doesn't](https://citedbyai.info/ga4-ai-assistant-channel): On 13 May 2026, Google Analytics 4 added a native AI Assistant channel to its Default Channel Group, automatically classifying referral traffic from ChatGPT, Gemini, Claude, and other AI assistants with an ai-assistant medium value. The channel reports how much AI-referred traffic a site receives and which platforms send it. It does not show how that traffic compares to competitors, which pages and paragraphs earn the underlying citations, whether AI platforms describe the brand accurately, or why the number is low when it is low. The rollout is gradual, not retroactive, and depends on referrer headers. GA4 measures the outcome of AI citation; a Citation Probability Score audit diagnoses the cause. Covers the four gaps, an outcome-versus-cause comparison, and the citable-content-to-citations-to-clicks chain. 4 FAQPage entries: what the channel is, why it shows no data, what it does not show, and how to improve AI Assistant traffic. - [Your First AI Citation Is Your Most Valuable](https://citedbyai.info/your-first-ai-citation): The strategic case for why the first citation a brand earns in any single AI engine (ChatGPT, Perplexity, Google AI Overviews, Gemini, or Claude) is disproportionately valuable. Three compounding mechanisms make first citations more durable than subsequent ones: citations persist into the next model retraining cycle, the citation-concentration gate is asymmetric (hard to pass, hard to dislodge), and the same citation graph feeds B2A agent discovery. Anchored on independent research: 15 domains capture 68% of all citations across the five major engines (5W AI Platform Citation Source Index 2026, 680M citations); only 11% of cited domains overlap between ChatGPT and Perplexity (ZipTie cross-platform study); 71% of cited sources appear on a single platform. Frames earning a first citation as a diagnostic problem before a content problem: map the citation graph in your category, score your own content for citability, target the high-citation domains for earned media. CBA's audit provides the inputs; outreach sits with the brand's PR team. 4 FAQPage entries on first-citation value, citation compounding, B2A discovery, and how to earn a first citation. - [AI Hallucination and Brand Compliance Risk: A Guide for Regulated Industries](https://citedbyai.info/ai-hallucination-brand-compliance-risk): A compliance-oriented guide for marketing and compliance teams in financial services, legal, healthcare, and professional services. Argues that sentiment monitoring (whether AI mentions a brand positively or negatively) is insufficient for regulated firms because it doesn't detect factual inaccuracies in AI descriptions. Key distinction: a brand can score well on sentiment while AI simultaneously misrepresents its regulatory permissions, service areas, staff credentials, or authorised activities. Evidence anchors: US courts imposed over $145,000 in AI hallucination sanctions in Q1 2026; researcher Damien Charlotin has catalogued over 1,353 AI hallucination cases globally; Ahrefs AI Benchmark Report found Gemini and Perplexity hallucinated in 37 to 39 percent of controlled experiment answers; EU AI Act high-risk deadline for financial services AI is August 2026; Sullivan and Cromwell apologised to a federal bankruptcy judge in April 2026 for AI hallucination events. Four industry risk cards cover financial services (regulatory status errors), legal (practice area and credential fabrication), healthcare (clinical scope and staff errors), and professional services (certification and scope misrepresentation). Explains the architectural gap between sentiment monitoring and hallucination detection. Includes an explicit scope note: CBA detects brand inaccuracies, not regulatory liability; the compliance interpretation sits with the firm's legal team. 4 FAQPage schema entries. - [Five Questions to Ask Any AI Visibility Platform Before You Sign](https://citedbyai.info/ai-visibility-platform-buyers-guide): Enterprise buyer's guide for teams shortlisting AI visibility platforms. Five diagnostic questions that separate tools with defensible methodology from those selling a black-box score: (1) Is the measurement methodology published? Most platforms describe their scoring as proprietary without publishing a framework, making the score impossible to audit or defend to a CFO. (2) Do you detect hallucinations or only count mentions? Most monitoring tools count brand appearances without checking whether AI descriptions are accurate. The Ahrefs AI Benchmark Report (May 2026) found Gemini and Perplexity hallucinated in 37 to 39 percent of answers in a controlled experiment. (3) Can you attribute visibility changes to revenue? Most platforms produce visibility scores without connecting to GA4-attributed revenue; the GA4 AI Assistant channel and 14.2% AI-referred conversion rate versus 2.8% organic baseline are the measurement chain that matters. (4) Do you cover all five platforms per-platform? Google Search Console as of June 2026 covers AI Overviews and AI Mode only; ZipTie research found only 11 percent of domains cited by ChatGPT are also cited by Perplexity for the same query. (5) Is the scoring framework independently validated? CPS framework validated by Perea Research May 2026, Princeton KDD GEO benchmarks, and Ahrefs Benchmark Report. Includes a visual scorecard rating each question as Most fail, Some pass, or All pass. Also includes a HowTo schema block with five steps matching the five questions. 5 FAQPage schema entries. - [Google Search Console AI Performance Reports: What They Show and What They Don't](https://citedbyai.info/gsc-ai-performance-reports): CBA's same-day response piece to Google's 3 June 2026 announcement of dedicated Search Generative AI performance reports in Search Console, rolled out first to a subset of UK website owners under pressure from the UK Competition and Markets Authority. The reports show impressions from AI Overviews, AI Mode, and generative AI features in Discover, across five dimensions: impressions, pages, countries, devices (Search only), and dates with hourly through monthly granularity. The reports do not show clicks, click-through rate, revenue, or any data from ChatGPT, Perplexity, Claude, or Microsoft Copilot. Google confirmed clicks are not yet available; a spokesperson stated Google will introduce additional metrics over time. Alongside the reports, Google is testing an opt-out toggle for UK sites to exclude themselves from AI Overviews, AI Mode, and Discover AI features, connected to the CMA's digital markets conduct requirements. The piece makes three arguments: (1) the new report covers Google's surfaces only, leaving ChatGPT, Perplexity, Claude, and Copilot unmeasured; (2) impressions without clicks is a partial picture that can't support revenue attribution; (3) a low impression number is a symptom requiring diagnosis that the GSC report itself cannot provide. Ends with CBA's five-platform audit as the natural next step. 4 FAQPage schema entries covering what the reports show, why the UK-first rollout, whether they cover ChatGPT or Perplexity, and what the opt-out toggle is. - [Bing Just Added Intents, Topics, Citation Share and Compare: What It Means for Your AI Visibility](https://citedbyai.info/bing-ai-visibility-insights-intents-topics-citation-share): CBA's response piece to Microsoft's 16 June 2026 announcement of four new AI visibility capabilities in the Bing Webmaster Tools AI Performance report (in preview, rolling out globally): Intents (classifies the grounding queries behind citations into categories such as Informational, Commercial, Navigational, Learn and Solve, Research, Creation, and Local), Topics (clusters related grounding queries into broader thematic groups), Citation Share (the percentage of citations attributed to your site out of all citations shown for a given grounding query), and Compare (overlays a previous time period onto the current view). The piece makes three arguments: (1) these features validate the intent, topic, share-of-voice, and trend framework CBA's audit has always used, formalising for Microsoft's surfaces the way CBA has scored AI visibility across all five platforms; (2) Citation Share deliberately hides competitors, with Microsoft stating explicitly that it 'does not expose competitor domains' and is 'designed as an observational metric, not a ranking system or a competitive scoreboard', which leaves head-to-head competitive benchmarking as the gap CBA fills by naming the specific domains winning the citations you are not; (3) the entire report covers Microsoft Copilot, Bing, and select partner experiences only, providing no native view of ChatGPT, Perplexity, Claude, or Google Gemini, mirroring the same single-platform boundary as Google Search Console's AI reports. Ends with CBA's five-platform audit with named head-to-head competitors as the way to fill both gaps. 3 FAQPage schema entries covering what Bing added, whether Citation Share shows competitors, and whether Bing Webmaster Tools shows ChatGPT/Perplexity/Claude/Gemini visibility. - [What the Ahrefs AI Benchmark Report Means for Your Brand](https://citedbyai.info/ahrefs-ai-benchmark-report): CBA's synthesis and practical framing of the Ahrefs AI Search Benchmark Report (Q4 2025 / Q1 2026), the most statistically robust AI search study published to date. Dataset: 146M SERPs, 730K AI responses, 75K brands, 174K cited pages, 76K websites. Covers three headline findings: (1) YouTube mentions were the single strongest correlate of AI brand visibility across 75K brands (r approximately 0.737), outperforming branded web mentions, domain rating, and backlinks by a wide margin; YouTube mention impressions were the second strongest factor. (2) Ahrefs researcher Mateusz Makosiewicz ran a misinformation experiment seeding three false stories about a fake brand; Gemini and Perplexity included misinformation in 37 to 39 percent of answers; ChatGPT-4 and 5 hallucinated in under 7 percent; Claude did not hallucinate but never surfaced the official website. (3) ChatGPT has 12 percent of Google's query volume but Google sends 190 times more traffic to websites; ChatGPT has a 96 percent lower CTR; AI-referred visitors convert at substantially higher rates than organic search traffic. Also covers: 58 percent CTR collapse at position 1 when an AIO is present; 28 percent of ChatGPT's most-cited pages have zero organic visibility; AI Mode and AI Overviews share only 13.7 percent URL overlap despite agreeing on conclusions 86 percent of the time; content length shows near-zero correlation with citation. Ends with three priority actions: audit YouTube and brand-mention footprint; check what AI platforms say about your brand; run the per-platform citation audit. 4 FAQPage schema entries. - [10 Best ASEO Tools 2026](https://citedbyai.info/best-aseo-tools): Ranked review of every major ASEO/GEO tool: Cited By AI®, Profound, AthenaHQ, Otterly, Peec AI, Scrunch AI, Conductor, SE Ranking, Bluefish AI. Pricing, platform coverage, block-level scoring, hallucination detection, and a four-gap capability comparison table. March 2026. - [Why AI Agents Can't Buy From Your Store](https://citedbyai.info/why-ai-agents-cant-buy-from-your-store): Why AI shopping agents ignore most ecommerce stores, what structured product data is, which fields determine agent inclusion (GTIN, brand schema, real-time availability, specific attributes), and how the Universal Commerce Protocol affects purchasability - [ASEO Audit Case Study: Local Garage](https://citedbyai.info/aseo-audit-case-study-local-garage): Real before-and-after ASEO audit data on an independent Audi/VW specialist in Alberta, Canada. Five audit runs, 29 content blocks, CPS 52.7 baseline, 2.2% Experienced Customer persona SOV. Published with client permission. - [Why Your GEO Score Is Wrong](https://citedbyai.info/why-your-geo-score-is-wrong): Why page-level GEO scores misrepresent AI citation performance, how RAG retrieval operates at the 134–167 word block level, and how block-level CPS auditing identifies the specific paragraphs being cited and skipped - [Citation Probability Score Framework](https://citedbyai.info/citation-probability-score-framework): Full five-pillar CPS framework article covering Content Structure, Fact Density, Answer Structure, Self-Containment, and Freshness Signals. Includes worked example, grade tiers, and open-source framework at github.com/citedbyai/cps-framework - [AI Crawler Simulator](https://citedbyai.info/ai-crawler-simulator): Free tool: enter any URL and choose an AI crawler identity to see exactly what GPTBot, ClaudeBot, PerplexityBot and 12 other AI bots can read on your site. Surfaces robots.txt blocks, llms.txt presence, and content visibility gaps in under 30 seconds - [AI Search Visibility](https://citedbyai.info/ai-search-visibility): Decision-stage guide explaining how Cited By AI® diagnoses why a brand is absent from AI answers and produces the specific fixes per page, per platform, with monthly SOV tracking - [ASEO Specialist](https://citedbyai.info/aseo-specialist): Why citability has replaced ranking, platform-by-platform citation behaviour, and the funnel gap most brands miss - [How to Choose an ASEO Provider](https://citedbyai.info/how-to-choose-an-aseo-provider): Evaluation framework for selecting an AI Search Engine Optimisation provider - [LLMs.txt Generator](https://citedbyai.info/llms-generator): Free tool: enter your domain, company name, and email to generate a production-ready llms.txt and llms-full.txt automatically. Crawls your sitemap, classifies URLs, writes contextual descriptions using Claude AI. Follows llmstxt.org spec. No signup required. ## Location - Gloucester, Gloucestershire, UK - Serving clients across the United Kingdom ## Contact - Web: https://citedbyai.info/#contact - LinkedIn: https://www.linkedin.com/in/citedbyai/ - X: https://x.com/citedbyai