Your buyers are asking ChatGPT about your category before they ever type your domain into a browser. Some of them will hear your name. Most won't. The difference is not whether you've done SEO — it's whether your brand has the property that decides who AI assistants name out loud. That property has a name now. AI citability — the measurable likelihood that ChatGPT, Claude, Perplexity, and Google AI Overviews will cite, mention, or recommend you by name when someone asks for the best in your category.
This guide is the canonical definition page, the 60-second diagnostic anyone can run before lunch, and the 30-prompt audit that turns a hunch into a baseline. It does not re-explain how AI engines extract chunks from a page — the generative engine optimization framework covers the input layer. This pillar is about the outcome — am I being recommended? — and what to do when the answer is no.
Key takeaways
- AI citability is a property, not a count. Visibility is "how many times did I appear?" Citability is "is my brand the kind of brand AI engines reach for in this category?" One is a number; the other is a structural fact.
- Three out of four is the bar. Run the 60-second self-test across ChatGPT, Claude, Perplexity, and Google AI Overviews. If your name doesn't come back from at least three, you have a measurable gap — and it almost always traces to entity absence or extraction friction.
- Buyers are already inside AI search. Industry surveys put AI-assistant use at roughly 9 in 10 B2B buyers during 2025 buying journeys, with ChatGPT the dominant tool. By the time a buyer hits your site, the recommendation has often already happened — or already not happened.
- The fix is five properties. Extractable answer blocks, primary-source citations, named-author authority, entity-consistent brand presence, and schema. Each maps to a graded sub-signal in the audit.
- Citation is unstable. Roughly 40–60% of cited domains change month to month. AI citability is a tracking job, not a one-time fix.
1. What "AI citability" actually means (and why we need a new word)
This is the canonical definition. If you backlink to it, this is the paragraph to point at.
The 30-second definition
AI citability is the measurable likelihood that an AI assistant will cite, mention, or recommend your brand by name when asked about your category. Three load-bearing parts: it is measurable, the recommendation is by name, and it happens in your category. Strip any one and you have a different concept — visibility, mentions, or generic AI exposure.
The existing vocabulary doesn't make this distinction. AI search visibility gets used as a count of appearances. AI mentions gets used for any time a brand name surfaces. AI citability names the underlying property — whether your brand is the kind AI engines reach for when someone asks the question your buyer is asking.
Why "citability" and not "visibility" or "mentions"
Visibility is a count. Mentions are events. Citability is a property — the structural fact that decides whether visibility and mentions happen at all. A page can be visible (indexed, crawled) without ever being cited. A brand can be mentioned without being recommended. AI citability is the upstream property that, when present, produces both.
AI citability vs SEO vs GEO
Three terms, one stack. Not interchangeable.
- SEO is the input layer for traditional search — the work that produces rankings.
- GEO (generative engine optimization) is the input layer for AI search — sometimes called AI search optimization — covering schema, llms.txt, robots.txt, page formatting, and off-site authority. That layer is covered in the generative engine optimization framework.
- AI citability is the outcome layer — the measurable property both produce. SEO optimizes for rankings. GEO optimizes for citations. AI citability is the score you get for the latter.
Stack of three: SEO is the foundation, GEO is the work, AI citability is the report card.
Industry surveys indicate roughly 94% of B2B buyers used at least one LLM during their 2025 buying journey, with about 63% of that AI-chatbot research happening on ChatGPT (surfaced via Deep Marketing; trail leads to a major analyst-firm report — verify primary before quoting in client decks). AI citability isn't a future metric. It's the channel where your buyers form their shortlist before they hit your site.
The audit angle: SEOGrade grades AI citability as one of nine categories. Naming the property is step one toward measuring it.
2. The 60-second AI citability self-test
Before you read another section, run the test.
Run the test yourself, right now
Open four tabs: ChatGPT, Claude, Perplexity, Google AI Overviews. Ask each the same question, written the way a buyer would type it: "What's the best [your category] for [your buyer]?" Then a second prompt: "What are the top [category] tools in 2026?"
Note who gets cited. Note where you appear — first mention, mid-paragraph, footer, or nowhere. Note your competitors who are cited and you aren't.
If your name doesn't come back from at least three of the four engines, you have an AI citability gap. If it doesn't come back from any, you're invisible in the channel where buyers spend roughly six weeks researching inside AI assistants before contacting a vendor for the first time (surfaced via Warmly; same provenance flag as the 94% figure).
What the result actually tells you
- You got cited. Your entity is established and at least one cited page is structurally extractable. The work shifts to share of voice and prompt-position.
- You didn't get cited. Either your entity isn't established or your content can't be extracted cleanly. Most sites have both problems.
Skipping the test is the third outcome — and the most expensive. 69% of buyers report AI assistants surfaced information that led them to choose a different vendor than originally planned (G2 — 2026 AI Search Insight Report). If your AI channel is unmonitored, the vendor switch already happened — you just don't know whose name replaced yours.
The quantitative version — 20–30 prompts across four engines, scored for citation rate, position, and sentiment — is the full 30-prompt audit protocol.
The audit angle: this is the manual version of what /audit automates across nine categories.
3. Why ChatGPT recommends your competitors instead of you
Three sources, three failure modes. The diagnosis is structural.
Three sources, three failure modes
The question buyers really type — how to get cited by ChatGPT — comes down to three sources. When ChatGPT decides who to recommend, it pulls from three places — the same three named in why ChatGPT recommends your competitors:
- Pretraining data — what was on the open web when the model was trained. Failure mode: you weren't there, or you were described differently than you'd describe yourself.
- Retrieval at query time — live web results pulled during the conversation. Failure mode: you're not in the underlying index (Bing for ChatGPT, Brave for Claude, Perplexity's own, Google for Gemini and AIO).
- Brand associations — entity-level signals from Wikipedia, Crunchbase, G2, LinkedIn, schema. Failure mode: your brand isn't an entity in the model's mental map of your category.
Most teams losing the citation war are losing on at least two of three. All three are fixable; the diagnosis tells you which to fix first.
The most common gap: entity absence
The single biggest reason an SMB or mid-market brand doesn't get recommended is that the model has no clean entity record for it. No Wikipedia page. No completed Wikidata entry. A thin Crunchbase profile. Inconsistent NAP across the open web — "Acme" on one site, "Acme Inc" on another, "Acme Solutions LLC" on the third. To the model, those are different entities. None are confidently linked to your category.
The second most common gap: extraction friction
If your homepage is a hero image and three feature bullets, the model has nothing to quote. That's the extraction-friction failure mode in one sentence. ChatGPT-cited pages contain structured data 71% of the time, and listicle formats account for roughly 43.8% of citation page types (surfaced via Wellows; verify the original Profound or Ahrefs Brand Radar study before client decks). Pages the model can extract get cited; pages that fight the model don't.
The full diagnostic — why each ai citability gap happens, what fixes it, in what order — lives in the diagnostic deep-dive.
The audit angle: SEOGrade's AI Citability category grades exactly this — entity presence, extraction-readiness, and citation share — as named sub-signals.
4. How AI answer engines actually choose what to cite
The mechanism is specific and tractable. The full walk-through of the input mechanics — query fanout, retrieval, synthesis — belongs to the generative engine optimization framework. This section covers what those mechanics mean for whether your brand shows up.
The pipeline, in one paragraph
Every AI answer engine does the same three things: it expands the prompt into 3–10 sub-queries, hits a search index for each one, and synthesizes the returned passages into a response with citations. The mechanics are the same; the personalities are different.
The five citation inputs every engine weighs
Strip away the engine-specific quirks and citation comes down to five inputs:
- Entity establishment. Is your brand a confident named entity in the category map? Wikipedia, Wikidata, Crunchbase, structured
Organizationschema feed this. - Extractability. Can the synthesizer lift a clean answer block without stitching across sections? FAQ schema, question-led H2s, and 40–60-word lead answers build it.
- Third-party corroboration. Do the platforms the model trusts (Reddit, YouTube, G2, LinkedIn, industry pubs) describe you the way you describe yourself?
- Recency. Content updated within 30 days receives roughly 3.2× more citations than older material (surfaced via Wellows; verify before quoting).
- Topical authority depth. Does your domain own a topic, or just touch one? Depth wins thinness.
The per-engine map at a glance
Same five inputs, different weighting:
| Engine | Dominant citation source | Freshness window | Live retrieval |
|---|---|---|---|
| ChatGPT | Wikipedia (~47.9% of top-10 sources); Reddit in >5% of all responses | 30–90 days | Routes through Bing |
| Perplexity | Reddit (~46.7% of top-10 citations) | ~6 months before steep decay | Always-on; proprietary index |
| Claude | Mixed; favors edited reference sites | 90 days | Routes through Brave Search |
| Google AI Overviews | Organic top 10 anchors ~99% of citations | News ~24h, evergreen ~90 days | Anchored to Google's index |
| Gemini | Google index plus Reddit; YouTube transcripts as 2026 development | Same as Google | Same surface; transcript-aware |
Source data: Search Engine Roundtable, Profound — AI Platform Citation Patterns.
Why citations are unstable
Profound's tracking shows roughly half of all cited domains change every month. A one-time audit ages out fast. The flip side: the citability gap you have today is actively recoverable if you start measuring.
The audit angle: each input maps to a graded signal in AI Citability. The audit reads them as a panel, not a checklist.
5. How to rank in ChatGPT specifically
ChatGPT is the engine doing most of the AI citation work for B2B buyers — about 900 million weekly active users as of February 2026, processing ~2 billion prompts a day (TechCrunch, DemandSage) — or, in the question buyers actually type, how to appear in ChatGPT. Four mechanics decide whether your brand shows up.
The Bing dependency
ChatGPT's live retrieval routes through Bing. If your site isn't indexed there — and a surprising number of mid-market sites still aren't — you're not in ChatGPT's live retrieval set. The fix is unglamorous: submit your sitemap to Bing Webmaster Tools, verify ownership, check coverage. Index in Bing first, then optimize for ChatGPT citability is the order most teams skip.
The Wikipedia and Reddit bias
ChatGPT cites Wikipedia in roughly 47.9% of top-10 cited sources and surfaces Reddit content in more than 5% of all responses. If your category has a Wikipedia article and you're not in it, the model has a primary source describing your category that doesn't mention you. The fix isn't manipulation — get notable enough to qualify for a Wikipedia article, participate honestly in the subreddit, get named by people who aren't you. Compounding takes months; the alternative takes forever.
GPTBot crawl rules
If GPTBot, OAI-SearchBot, or ChatGPT-User can't reach your pages, no downstream optimization saves you. The minimum-viable robots.txt for ChatGPT citability:
User-agent: GPTBot
Allow: /
User-agent: OAI-SearchBot
Allow: /
User-agent: ChatGPT-User
Allow: /
This is the floor, not the ceiling — the full bot-access list lives in section 10. Decide on purpose, not by default.
The brand-name pattern
ChatGPT recommends by entity, not URL. If your brand is named identically across Wikipedia, Wikidata, Crunchbase, G2, LinkedIn, and your own site, the model has a confident entity to recommend. If you're "Acme" in three places and "Acme Inc" in two more, the model has ambiguity to resolve — and resolves it by recommending your competitor instead.
For the prompt-by-prompt fix order, the diagnostic deep-dive goes signal by signal. For the marketing-level framing, see the AI SEO approach.
The audit angle: GPTBot access, Bing index presence, and entity consistency all show up as named ChatGPT-citability signals.
6. How to rank in Perplexity
Perplexity is the engine where Reddit savvy decides AI citability outcomes.
Reddit is the engine room
Perplexity sources roughly 46.7% of top-10 citations from Reddit (Profound). That's a structural choice, not a quirk. The playbook is honest participation — pick the two or three subreddits where your buyers spend time, answer questions without dropping links unless asked, get named by other users in threads that aren't yours. Astroturfing gets caught and torched; consistent presence compounds.
The 6-month freshness decay
Perplexity de-weights content older than roughly six months. Last-updated stamps, quarterly refresh cycles, and a dateModified field in your Article schema all feed the engine's freshness check. A 2024 post that was great in 2024 is, by Perplexity's logic, no longer cite-worthy in 2026.
Sonar = Google rankings + Reddit + schema
Perplexity's retrieval shorthand: it weighs Google-style organic rank, Reddit signal, and structured data. Win all three, citation is almost automatic. Win two, it's likely. Win one, you're losing to a competitor winning two.
The audit angle: Reddit citation share, freshness, and Sonar-style triangulation show up as Perplexity-citability proxies.
7. How to rank in Claude
Claude is the strictest engine — and the one most B2B teams under-invest in for AI citability.
The Brave Search dependency
Claude's live retrieval routes through Brave Search. Available studies show roughly 86.7% citation overlap with Brave's index. Brave is smaller than Google or Bing and genuinely under-coupled from the rest of the SEO stack. Submit to Brave is a bullet most SEO checklists don't have. It belongs on yours.
The 90-day freshness rule
Claude's freshness window is tighter than Perplexity's. The heuristic: a page last updated within 90 days is citation-eligible; older pages get deweighted. The fix is operational — a quarterly refresh cycle on every pillar post, with real edits (not dateModified bumps). Claude is increasingly good at catching faked freshness.
Free audit
Grade your site free
See how you score across all 9 categories — in roughly 60 seconds. No signup.
Schema is non-negotiable
Claude punishes missing schema hardest. Pages without FAQPage, Article, or HowTo markup see materially lower citation rates than marked-up equivalents. JSON-LD only. Validate every block in Google's Rich Results Test — invalid schema is worse than no schema.
The audit angle: Brave indexing, schema validation, and 90-day freshness show up as named Claude-citability signals.
8. How to rank in Google AI Overviews and Gemini
This is the cluster boundary section. The on-page mechanics for getting featured in AI Overviews — schema implementation, passage formatting, comparison-table structure — are owned by the on-page mechanics for getting featured in AI Overviews. This pillar covers what it takes for your brand to be the cited entity.
The top-10 organic dependency
Roughly 99% of AI Overview citations come from the organic top 10, with about 70% from the broader top 100. AIO is a re-ranking layer on top of organic results, not a replacement. AI citability and traditional SEO can't be uncoupled — a site invisible in Google's organic SERP is invisible in AIO almost by definition.
Schema as the AIO multiplier
Pages with FAQPage and HowTo schema show roughly 2–2.5× higher AIO citation likelihood than unmarked equivalents. Schema doesn't cause citations; it makes the page extractable enough that citation can happen at all. Brands competing for the same top-10 organic slot win or lose AIO citations on schema parity.
Comparison tables and benchmark data
AIO favors passages in the 134–167-word sweet spot that contain comparison tables or benchmark data. Every category-defining page should have at least one comparison block — your product against the alternatives, with the columns that matter, scored honestly.
Gemini and the YouTube development
Gemini draws from Google's index plus Reddit, with one 2026 development worth flagging — YouTube transcripts are increasingly the citation source for product-explainer queries. Gemini ingests transcripts, not pixels. A clean explainer video with chapter markers and a real transcript creates a citation surface that didn't exist on your domain.
The audit angle: organic rank, schema validation, and brand entity recognizability are all graded inside AI Citability — and the audit reads them together, because in AIO they only matter together.
9. The 5 properties of cite-worthy content
If AI citability is the property that produces citations, here are the five sub-properties that produce citability.
Property 1: Extractable answer blocks
Lead every major H2 with a 40–60 word answer in plain prose. Question-led headings — written the way a buyer would type the query — match how the synthesizer maps prompts to passages.
Property 2: Primary-source citations
Link to original studies, papers, and primary data — not roundups that reference them. Citation-of-citations decays through the chain; the model trusts the source closest to the receipt.
Property 3: Named-author authority
Bylines by named experts, with bios, credentials, and consistent presence across the open web. E-E-A-T at the author level, not just the site level. As Lily Ray, VP SEO Strategy at Amsive, keeps saying: AI search amplifies E-E-A-T, it doesn't replace it.
Property 4: Entity-consistent brand presence
Brand name, category description, and value proposition described identically across Wikipedia, Wikidata, Crunchbase, G2, LinkedIn, and your homepage. If they describe five different companies, the model picks one — and "the right one" isn't guaranteed.
Property 5: Schema that maps the page to AI crawlers
FAQPage, HowTo, Article, Organization. JSON-LD only. Pages with schema get cited at meaningful multiples over pages without; pages with invalid schema get cited less than pages with no schema at all.
| Property | 60-second test |
|---|---|
| Extractable answer blocks | Can the first paragraph under each H2 stand alone? |
| Primary-source citations | Of inline citations, how many link to a primary vs a roundup? |
| Named-author authority | Click the byline — real, consistent named author? |
| Entity-consistent presence | Search your name across Wikipedia / Crunchbase / G2 — descriptions match? |
| Schema mapping the page | Run the URL through Google's Rich Results Test; any errors? |
The Princeton/IIT Delhi GEO paper is the load-bearing receipt for AI citability. Adding statistics, citations, and quotations to a page lifted source visibility by up to 40% in generative-engine responses (arxiv.org/abs/2311.09735) — the highest-impact tactic the authors studied. Same techniques produce GEO visibility and AI citability. Same lever, two outcomes.
The audit angle: each of the five properties is a graded sub-signal inside AI Citability. The full operational checklist lives in the framework deep-dive (cluster footer).
10. Technical setup: schema, robots.txt, llms.txt
The configuration layer in three short rules.
Letting the right bots in
The minimum AI-bot allow-list goes well beyond GPTBot. Order them in your robots.txt:
GPTBot,OAI-SearchBot,ChatGPT-User— ChatGPT's training, search, and conversational fetches.ClaudeBot— Anthropic's crawler.PerplexityBot— Perplexity's crawler.Google-Extended— Gemini training opt-in. Distinct fromGooglebot; blockingGoogle-Extendeddoesn't affect Google Search, only Gemini.
Decide each one deliberately. Blocking them all is a citability-killing default; allowing them all is a licensing question worth answering on purpose.
llms.txt: probably skip it
llms.txt is a proposed standard — a markdown manifest at your site root listing canonical pages you'd like LLMs to prioritize. Anthropic and Mintlify publish one. Google's John Mueller has stated no AI system actually consumes it, and available studies indicate roughly 94.9% of llms.txt requests come from Googlebot rather than from any AI crawler. Ship it if your site already has clean canonicals — and skip it if you're triaging. It's not where the citability problem lives.
Schema priorities for AI citability
Four types do most of the work: FAQPage, HowTo, Article, Organization. FAQPage carries the largest citation-likelihood lift in available data — so if you ship one schema type first, that's the one. Validate every block. Invalid schema signals "this site doesn't maintain its technical surface" — the inverse of what schema is supposed to communicate.
The audit angle: bot access, schema validation, and structured-data completeness are graded directly. The audit names the gap before you guess.
11. How to measure AI citability
AI citability measurement breaks into four metrics, two protocols, and a tools landscape.
The four metrics that matter
- Citation rate — across a fixed prompt set, what percentage of responses cite your brand?
- Share of voice — of all citations on your prompt set, what percentage are you vs each competitor?
- Sentiment — when you're cited, is the framing positive, neutral, or negative? Negative citations aren't wins.
- Prompt-position — first mention, mid-paragraph, footer? Position predicts click-through and recall.
Manual measurement: the 30-prompt protocol
Pick 20–30 prompts a real buyer would type — mix head, comparison, use-case, and problem-driven shapes. Run each through ChatGPT, Claude, Perplexity, and Google AI Overviews. Score against the four metrics, one row per prompt-engine pair. That's your ai citability baseline. Re-run monthly. This is the 30-prompt audit protocol the audit funnel runs on.
Automated tools, briefly
The current landscape: Profound (citation tracking and drift data), Otterly (multi-engine prompt monitoring), Peec AI (referral analytics), Brandlight (brand-mention monitoring), Ahrefs Brand Radar, Semrush AI features (AIO tracking). One sentence each because a head-to-head shootout belongs in its own post.
Aleyda Solís, who co-contributed to Microsoft's February 2026 AI marketers' guide, makes the sharper point: tools that only show where your brand appears in AI answers aren't enough. AI citability insights have to integrate into the broader SEO strategy or they end up as a side dashboard nobody acts on. Roughly 67% of digital marketers say GEO and AI tracking is more complex than traditional SEO measurement (HubSpot; verify primary before quoting).
The audit angle: SEOGrade's audit produces citation rate and share of voice automatically inside AI Citability — weekly snapshots, not point-in-time guesses.
12. Audit your AI citability — the 30-prompt protocol
The full deep-dive — prompt templates, scoring rubric, spreadsheet format — lives in the prompt-by-prompt audit protocol. Here's the three-step shape.
Step 1: pick 20–30 prompts a real buyer would type
Four prompt shapes:
- Head — "best [your category] for [buyer type]". Where category-defining citation lives.
- Comparison — "[competitor] alternatives", "[competitor] vs [competitor]". The most commercially loaded shape.
- Use-case — "[your category] for [vertical]". Mid-funnel.
- Problem-driven — "how do I [pain point]?". Top-funnel.
At least five prompts per shape. Twenty is the floor; thirty is better.
Step 2: run across four engines, score each
For each prompt-engine pair (20 prompts × 4 engines = 80 runs), record: citation appears (yes/no), position, sentiment, and competitor citation rate. Sixty to ninety minutes for a 20-prompt panel. The output is a baseline number, not a feeling.
Step 3: triage by gap type
The shape of your AI citability gap tells you what to fix first:
- Entity gap — cited rarely or not at all across most engines. Wikipedia / Wikidata / Crunchbase / consistent NAP work.
- Extractability gap — indexed and ranking, not cited. Schema, content rewrite, 40–60 word lead answers.
- Recency gap — used to be cited; slipping. Quarterly refresh cycle with substantive edits.
- Corroboration gap — cited only on your own pages, never via Reddit / Wikipedia / G2. Third-party content strategy.
Most sites have two of the four. The audit names which two — and which to fix first.
Run the manual version yourself with a spreadsheet and three hours. Or grade your site free in 60 seconds and let the AI Citability category run the prompt panel, the scoring, and the gap triage. Either way, act on your audit findings is the step nobody can skip.
The audit angle: this is the audit angle. The audit automates the protocol across all nine categories.
13. How AI citability fits with traditional SEO and GEO
Three layers, three jobs.
- SEO is the foundation. Entity, schema, crawlability, content quality. The SEO audit checklist 2026 covers what stays the same.
- GEO is the input layer for AI search. Page formatting, robots.txt, llms.txt, schema, off-site authority. The generative engine optimization framework is the playbook.
- AI citability is the outcome. Citation rate, share of voice, sentiment, prompt-position — the score, not the work.
The mistake most teams make is treating AI citability as a side metric — separate dashboard, separate consultant, separate budget line — instead of one of nine grades on the same report card. SEOGrade's framework grades SEO, GEO, and AI Citability as three distinct categories, alongside Crawlability, Technical, On-Page, Content & E-E-A-T, Authority, pSEO, and Local. Browse all 9 audit categories for the full breakdown. For the on-page mechanics, that's the generative engine optimization framework's territory. We measure the outcome here.
14. Frequently asked questions
What is AI citability?
AI citability is the measurable likelihood that an AI assistant will cite, mention, or recommend your brand by name when asked about your category. It's the outcome SEO and GEO produce — the property that makes citations happen at all.
How do I check if ChatGPT recommends my brand?
Open ChatGPT, Claude, Perplexity, and Google AI Overviews. Ask each: "What's the best [your category] for [your buyer]?" If your name doesn't come back from at least three of the four, you have a citability gap. The full version is the 30-prompt audit protocol.
How is AI citability different from AI search visibility?
Visibility is a count — how many times did your brand appear? Citability is a property — is your brand the kind AI engines reach for in your category? Visibility measures events; citability measures the structural fact that produces them.
Does ChatGPT cite my website directly?
Rarely as URLs; usually as a brand name. ChatGPT recommends by entity, not URL. Entity consistency matters more than URL-level optimization for citability.
How often does ChatGPT update its sources?
The training data lags by months. Live retrieval via Bing is real-time. Profound's drift data shows roughly half of cited domains change month to month — treat it as a moving baseline.
Do I need llms.txt?
Probably not yet. Google's John Mueller has stated no AI system uses it; most llms.txt requests come from Googlebot, not AI crawlers. Ship it if your site already has clean canonicals; skip it if you're triaging.
How long does AI citability take to improve?
Early signals show up in 4–8 weeks — citation appearances on softer prompts, entity-record updates landing, schema validation lifting AIO eligibility. Stable citation patterns take 3–6 months.
How do I track my brand mentions in AI?
Two paths. The manual 30-prompt protocol gives a real baseline in three hours. Automated tools (Profound, Otterly, Peec AI, Brandlight, Ahrefs Brand Radar) handle ongoing tracking.
AI citability isn't a soft metric — it's whether the buyers spending six weeks inside ChatGPT find your name when they ask. Grade your site free in 60 seconds. Nine categories, one score, AI Citability included.
More in this series
Supporting articles in the AI Citability cluster publish on a rolling schedule:
- The AI Citability Framework: 5 Properties of Cite-Worthy Content
- Perplexity SEO: How to Optimize for the Reddit-Heavy AI Engine
- Claude SEO: How to Get Cited by Claude
- Google AI Overviews and Gemini: Ranking Factors That Earn Citations
- AI Search Visibility Tools: The 2026 Landscape Compared
- How to Track ChatGPT Mentions of Your Brand
- LLM SEO: The Technical Setup for GPTBot, ClaudeBot, and PerplexityBot
Each will be linked inline once published.