
Henry Nguyen
Co-Founder, CiteFlow
Henry leads product and engineering at CiteFlow. The guides on this blog come from the same work the product does every week: asking ChatGPT, Perplexity, Gemini and Claude the questions buyers ask, and recording which pages the answers cite. Each guide links the sources it draws on.
How these guides are written
Guides are drafted with AI assistance and published by CiteFlow. Sources are linked in the text.
Guides by Henry
- How often AI gets your company's basic facts wrongAI assistants routinely misstate pricing, founding year, HQ, and integrations. Here's where the errors cluster and what structural fixes close the gap.
- What % of sites block AI crawlers in 2026, and who unblockedRoughly a quarter of top sites block major AI crawlers in 2026, but the internet-wide rate is far lower — and news publishers are moving to allowlists.
- Gemini vs Google AI Overviews: same company, different sourcesGemini and Google AI Overviews share a parent company but cite almost entirely different domains. Here's why, and which surface to optimise for first.
- AI citation rate benchmarks: SaaS vs ecommerce vs local servicesPublished studies show citation rates and dominant source types diverge sharply by vertical — local services lag Google Maps 2:1, while SaaS reviews sites and vendor domains do the heavy lifting.
- AI citation volatility: a 90-day tracking study of weekly churnWe tracked a fixed prompt set across engines for 90 days. Citations churn 40-75% weekly depending on the engine, which changes how you should measure GEO progress.
- BreadcrumbList schema: minor for Google, useful for AI topic mappingBreadcrumbList schema barely moves Google's needle anymore, but it still hands AI systems a clean read on your site's hierarchy. Here's the honest version.
- SoftwareApplication schema for SaaS: what AI assistants actually pullA field-by-field look at SoftwareApplication schema for SaaS products, what large language models tend to extract, and which fields to leave out rather than fake.
- Schema markup validation vs AI usefulness: why valid isn't enoughA green check in the Rich Results Test or Schema.org validator confirms syntax. It says nothing about whether AI assistants trust or use your markup.
- How to Check if ChatGPT Is Citing Your WebsiteYou can check whether AI assistants cite your site in about 15 minutes. Here is the exact routine: the questions to ask, where each assistant shows its sources, and how to keep results honest.
- Article schema done right: dates, author, and what breaks itA minimal valid Article schema block for AI citations, plus the three mistakes — bad dates, string authors, missing publisher logo — that get it ignored.
- FAQ schema in 2026: Google killed the rich result, is it still worth it?Google quietly removed FAQ rich results in May 2026. Here's whether FAQPage markup still earns citations in ChatGPT, Perplexity, and AI Overviews.
- Do Core Web Vitals matter for AI search? What the evidence saysAI crawlers fetch HTML, not render pages — so Core Web Vitals work as a gate for crawl completion and rankings, not a direct citation lever.
- Rate limiting AI crawlers: how 429s quietly kill citationsPerplexityBot and GPTBot don't retry forever. Aggressive rate limits and "anti-scraping" presets can throttle them into giving up before they ever index your content.
- Shopify and AI crawlers: what merchants can and can't controlShopify locks down robots.txt and theme rendering more than most platforms. Here's exactly what levers merchants have over AI crawler access and citations.
- Adding llms.txt to WordPress: manual, plugin, and what to includeThree ways to add llms.txt to WordPress — static file, functions.php route, or plugin — plus what to include, exclude, and how to keep it current.
- Cloudflare blocking AI crawlers? Why WAFs hide you from ChatGPTYour robots.txt probably allows GPTBot. Your WAF probably doesn't. Here's how to test for edge-level AI crawler blocks and fix them.
- Stale content and AI search: why old posts hurt citations sitewideOutdated pages don't just lose rankings quietly — they can drag down how often AI engines trust and cite your entire domain.
- The About page as an entity document: what to put on it for AIAI engines read About pages to resolve who you are, not to be convinced. Here's the checklist for writing one as an entity document.
- Making your pricing page readable to AI assistantsAI assistants hallucinate pricing when tiers live behind JS toggles. Here's how to render plain-text prices, add Offer schema, and handle enterprise tiers.
- Why AI engines love X vs Y pages (and how to write one)Comparison queries dominate buyer prompts. Here's how to structure an X vs Y page — neutral table first — that AI engines actually trust enough to cite.
- Competitor citation gap analysis: find the prompts where you loseA step-by-step method for finding the exact buyer prompts where AI engines cite competitors instead of you, and prioritizing fixes by revenue intent.
- Local business AI search: will ChatGPT recommend the plumber down the street?ChatGPT, Perplexity, and Gemini don't read your Google Business Profile the way you think. Here's where each one actually pulls local data from.
- AI search market share in 2026: the numbers, and the vendor hypeThe real 2026 data on AI search share — query volume, referral traffic, and click erosion — with vendor bias flagged and a clear threshold for when to act.
- We asked AI 450 times which project management tool to buy. Notion never came up.An original study of 450 AI answers across ChatGPT, Claude and Perplexity. A review blog got cited more than any product, Notion and Linear were never mentioned once, and the three engines agreed on almost nothing.
- Ahrefs Brand Radar vs CiteFlow: track AI visibility, or fix it?Brand Radar measures how often AI answers mention your brand. CiteFlow audits the site behind those mentions and drafts the fixes. They solve different halves of the same problem. Here is how to tell which half you are stuck on.
- AI citation traffic value: the zero-click math explainedAI citations send fewer clicks per impression than a #1 Google ranking, but the intent gap changes the economics. Here is how to model the break-even honestly.
- AI citation vs brand mention: why the two metrics divergeAn AI mention and an AI citation look similar in a dashboard but are earned by different signals and produce different business outcomes. Here is how to tell them apart.
- Which AI search engine your buyers actually use, by intentNot every AI engine attracts the same buyer at the same moment. Match your GEO strategy to the engine your target intent actually lives on.
- ChatGPT vs Google AI Overviews: how much do their citations actually overlap?Citation overlap between ChatGPT and Google AI Overviews is far lower than most teams assume — and the gap reveals two very different definitions of authority.
- Content freshness and Perplexity citations: how recent is recent enough?Perplexity's citation algorithm leans heavily on recency. Here are the actual windows, refresh cadences, and timestamp signals that decide whether you get picked.
- .com vs other TLDs: does your domain extension affect AI citation odds?Roughly 80% of AI citations point to .com domains. Here's why that's mostly correlation, and what it means if you run a .io or .ai brand.
- Reddit's share of Perplexity citations: what the 46.7% number actually meansThe widely cited 46.7% Reddit-Perplexity figure is a top-10 share, not total volume. Here's what the underlying data actually says.
- Product schema that AI shopping assistants actually readA practitioner's guide to Product JSON-LD for ChatGPT Shopping, Gemini, and Perplexity — with a real example and the fields that actually matter.
- sameAs vs knowsAbout: which Schema.org property whensameAs handles entity identity; knowsAbout declares topical expertise. Here's when to use each, JSON-LD examples, and the misuse patterns that break AI citations.
- Entity blindness: why ChatGPT sometimes ignores your schemaSchema markup alone won't force an LLM to recognize your brand as an entity. Here's what corroboration and cross-source consistency actually do.
- Organization schema + sameAs: the entity backbone for AI citationsA practical guide to wiring Organization schema and 8-12 sameAs URLs so ChatGPT, Perplexity, and Gemini attribute citations to the right entity.
- Canonical tags and duplicate content in AI search: what actually worksHow AI engines handle canonicals, syndication, and duplicate content — and why attribution dilution is the real risk for GEO in 2026.
- llms.txt in 2026: low-cost bet or waste of time?llms.txt has ~10% adoption and zero commitments from major LLMs. Here's the honest cost/benefit for 2026, and who's actually reading the file today.
- SSR vs CSR for AI crawlers: which bots actually execute your JavaScriptMost AI crawlers fetch HTML and stop. If your content, schema, or metadata depends on JavaScript, ChatGPT, Claude, and Perplexity cannot see it.
- Monitor AI crawler server logs: GPTBot, ClaudeBot, PerplexityBotA practitioner's guide to spotting GPTBot, ClaudeBot, and PerplexityBot in raw server logs — user agents, filters, and tooling that GA4 can't give you.
- Building topical authority clusters for AI searchA practitioner's guide to hub-and-spoke topical clusters that signal depth to LLMs and compound into citations from ChatGPT, Perplexity, and Gemini.
- Evidence-dense writing for GEO: swap marketing copy for citable factsLLMs cite verifiable claims, not adjectives. Here's how to rewrite marketing copy into evidence-dense passages that engines will actually quote.
- Answer-block formatting: structure content so AI can extract itHow to format answer blocks so ChatGPT, Perplexity, and AI Overviews can lift your content cleanly — with before/after markup examples.
- How to earn Perplexity citations through Reddit (without getting banned)A practitioner's playbook for earning Perplexity citations from Reddit through authentic participation, smart subreddit selection, and AMAs.
- Why AI cites Reddit and Wikipedia more than your homepageReddit and Wikipedia dominate AI citations while brand pages get scraps. Here's the structural reason — and what to do about it without becoming either.
- Do backlinks matter for AI citations? An honest 2026 answerBacklinks still influence which pages LLMs see, but co-citations and unlinked mentions now do heavier lifting. Here's what the recent studies actually show.
- The 7 signals LLMs use to pick which brands to citeSeven concrete signals — from brand search volume to co-citation — determine which brands LLMs surface. Here's what each one is and the action it maps to.
- GEO vs SEO in 2026: what actually changes (and what doesn't)GEO and SEO share infrastructure but diverge on signals, measurement, and content shape. Here's what carries over and what you have to rebuild.
- Robots.txt AI crawler audit: a 5-minute checklistA practitioner's 5-minute audit for GPTBot, ClaudeBot, PerplexityBot, and Google-Extended — with the exact directives and verification steps.
- How long until ChatGPT cites a new domain? Realistic timelinesBenchmarks from Profound, Semrush, and AuthorityTech on how fast ChatGPT cites new pages — plus the factors that compress or extend the timeline.