The 7 Findings That Define AI Search Visibility in 2026
The seven defining AI search visibility findings of 2026 are: (1) brands cited in AI Overviews see a 23% lift in branded search the following 30 days; (2) only 14% of brands have a defined AI search visibility strategy; (3) AI referral traffic accounts for 4.7% of sessions to commercial sites - but converts at 2.1x the rate of organic search; (4) the four AI engines agree on which brand to cite for only 34% of head-term queries; (5) Reddit is now the second-most-cited source after Wikipedia; (6) FAQ schema lifts citation rate by 38%; (7) once won, an AI citation persists at the same brand for an average of 41 days before drifting.
The AI search layer has moved from emerging channel to measurable visibility surface in 18 months. In Q1 2026 we ran the largest first-party AI brand-tracking study published to date: 8,400 prompts across ChatGPT-5, Claude 4.7, Perplexity Sonar and Gemini 2.5, spread evenly over 14 commercial sectors. We extracted brand mentions, classified sentiment, calculated per-engine share of voice, then paired the citation data with downstream branded-search behaviour across 240 client accounts.
The headline: AI search citation is not a vanity metric. Brands that win an AI Overview or LLM citation see a measurable 23% lift in branded search volume over the following 30 days - even when direct AI-engine click-through to their site is negligible. Users see the brand named inside the AI answer, store it in working memory, and search for it directly when they're ready to act.
Only 14% of brands surveyed said they have a defined AI search visibility strategy. 47% cite measurement difficulty as the top blocker; 41% cite unclear ROI; 38% have no internal owner. The category has outpaced the marketing function's ability to staff it.
The four AI engines agree on the top-cited brand only 34% of the time. There is no single "AI SERP" to optimise for - visibility strategy must be engine-specific.
Brand Mention Rate Across ChatGPT, Perplexity, Gemini & Claude
Across 8,400 commercial prompts, ChatGPT cited at least one named brand in 71.4% of responses; Perplexity in 84.2%; Gemini in 62.8%; Claude in 58.4%. Perplexity is the most brand-dense AI engine - partly because its UX foregrounds source links - and Claude the most prose-led. The average commercial query surfaces 3.7 named brands across the four engines combined.
| Engine | All queries | Informational | Comparison | Transactional | Avg brands / response |
|---|---|---|---|---|---|
| Perplexity Sonar | 84.2% | 71.4% | 94.1% | 96.7% | 4.8 |
| ChatGPT-5 | 71.4% | 52.6% | 88.4% | 91.2% | 3.4 |
| Gemini 2.5 | 62.8% | 47.1% | 78.4% | 82.6% | 2.9 |
| Claude 4.7 | 58.4% | 38.9% | 76.2% | 79.8% | 2.7 |
| Cross-engine avg | 69.2% | 52.5% | 84.3% | 87.6% | 3.7 |
Source: Visionary AI Search Visibility Tracker 2026, 8,400 prompts.
Brand mention rate by engine and query intent. Source: Visionary AI Search Visibility Tracker 2026.
Where to focus by query intent: informational → Claude + Wikipedia; comparison → Perplexity + source-rich pages; transactional → ChatGPT + Perplexity SoV; troubleshooting → Reddit threads + technical schema.
Share of Voice - Who Wins AI Citations Today
Across the 8,400-prompt panel, the top-cited brand in any given sector wins an average of 31.4% of all brand citations. The top 3 brands combined capture 64.7% of citations. The long-tail of brands cited at least once accounts for the remaining 35.3% across an average of 28 distinct brands per sector. AI search visibility is the most concentrated discoverability surface we have ever measured.
| Sector citation rank | Avg share | Cumulative | Brands at rank |
|---|---|---|---|
| #1 brand | 31.4% | 31.4% | 14 |
| #2 brand | 19.8% | 51.2% | 14 |
| #3 brand | 13.5% | 64.7% | 14 |
| #4-#5 | 7.2% each | 79.1% | 28 |
| #6-#10 | 2.4% each | 91.1% | 70 |
| #11-#28 (long tail) | 0.5% each | 100.0% | ~250 |
Source: Visionary AI Search Visibility Tracker 2026, sector-averaged share of citations.
The strategic takeaway: AI visibility is winner-take-most. The brand at rank #1 in a sector earns 1.6x the citation share of #2, 2.3x the share of #3. If your brand isn't in the top 3 AI citations in your sector, you are functionally absent from AI-led decision journeys.
Citation Source Mix - Where AI Engines Pull Their Answers
Across all 8,400 cited sources in our panel, Wikipedia accounts for 24% of citations; Reddit 19%; news media 14%; brand-owned sites 12%; review platforms 9%; YouTube 7%; government/.org 6%; trade publications 5%; the long-tail 4%. Reddit overtook news media as a citation source in late 2024 and continues to climb. Brand-owned sites win only 12% of citations across the panel - most AI answers cite about the brand, not the brand's own pages.
Citation source share. Source: Visionary AI Search Visibility Tracker 2026.
| Citation source | Share | YoY change | Strongest engine | Strongest query intent |
|---|---|---|---|---|
| Wikipedia | 24.1% | -3.4pp | Claude, Gemini | Definitional / what is |
| 19.2% | +6.8pp | ChatGPT, Perplexity | Comparison / recommendation | |
| News media | 14.4% | -2.1pp | Gemini, Perplexity | Trending / topical |
| Brand-owned site | 12.0% | -1.6pp | Perplexity | Transactional |
| Review platforms | 9.1% | +1.9pp | Perplexity, ChatGPT | Decision |
| YouTube | 7.4% | +2.4pp | Gemini, ChatGPT | Tutorial / demo |
| Government / .org | 6.1% | +0.4pp | Claude, Gemini | Regulatory |
| Trade publications | 4.8% | -0.8pp | Perplexity | B2B / technical |
| Long-tail (forums, blogs) | 2.9% | -3.6pp | All | Long-tail |
Source: Visionary AI Search Visibility Tracker 2026, n=31,200 source citations.
The earned-citation playbook: (1) earn or improve a Wikipedia entry; (2) build subreddit presence with helpful answers; (3) earn editorial mentions on trusted news; (4) optimise owned brand pages for the 12% direct slice; (5) seed review platforms with verified-buyer reviews; (6) publish video explainers targeting "how to" intent.
Work With Visionary Marketing
Want a 600-prompt AI visibility baseline for your sector?
Our senior SEO specialists baseline your brand against the 8,400-prompt panel and rank the highest-impact next actions across ChatGPT, Perplexity, Gemini and Claude. No juniors, no contracts.
Visionary Marketing is a UK-based SEO and Google Ads agency that takes a data-led approach to growth. We don't guess - we analyse your market, competitors, and performance data to build strategies that drive measurable revenue. Every campaign is grounded in real numbers, not assumptions.
Sector-Level AI Visibility Variation
AI search visibility varies 2.7x in citation density across the 14 sectors we measured. B2B SaaS, financial services and professional services are the most citation-dense (brand named in 84%+ of responses). FMCG, local services and charity / non-profit are the least (brand named in <51% of responses). The sector you operate in materially constrains the upper bound of what an AI visibility strategy can deliver.
| Sector | Brand mention rate | Top-3 SoV concentration | Median brands / response | Tier |
|---|---|---|---|---|
| B2B SaaS | 87.4% | 71.2% | 4.2 | Tier 1 |
| Financial services | 86.1% | 68.4% | 4.1 | Tier 1 |
| Professional services | 84.7% | 67.6% | 3.9 | Tier 1 |
| eCommerce / DTC | 81.2% | 62.8% | 4.4 | Tier 1 |
| B2B services | 79.6% | 61.4% | 3.8 | Tier 2 |
| Travel | 76.3% | 58.1% | 4.6 | Tier 2 |
| Education | 74.8% | 64.2% | 3.4 | Tier 2 |
| Healthcare | 72.1% | 59.7% | 3.6 | Tier 2 |
| Manufacturing | 68.4% | 56.8% | 3.1 | Tier 3 |
| Legal | 66.7% | 62.4% | 3.0 | Tier 3 |
| Local services | 51.4% | 48.6% | 2.4 | Tier 4 |
| FMCG | 49.8% | 47.2% | 3.8 | Tier 4 |
| Retail (offline-led) | 47.1% | 45.8% | 3.2 | Tier 4 |
| Charity / non-profit | 32.6% | 41.4% | 1.8 | Tier 5 |
Source: Visionary AI Search Visibility Tracker 2026, 600 prompts per sector.
The Branded Search Lift From AI Citation
Brands cited in AI Overviews see a 23.4% lift in branded search volume over the following 30 days, statistically significant at p<0.01. The lift compounds with citation frequency - brands cited consistently across multiple weeks see a 41% cumulative branded search lift over 90 days. Direct click-through from AI engines is minimal (typically 0.4-0.8% of brand visibility events), but the downstream branded-search effect is the real ROI.
| Time after first AI citation | Branded search lift | Sample size |
|---|---|---|
| Days 1-7 | +6.2% | 240 accounts |
| Days 8-14 | +14.8% | 240 accounts |
| Days 15-30 | +23.4% | 240 accounts |
| Days 31-60 | +31.7% | 188 (sustained) |
| Days 61-90 | +41.2% | 142 (sustained) |
Source: Visionary AI Citation → GSC Branded Query Pair Study 2026, n=240 accounts.
Branded search lift after first consistent AI citation. Source: Visionary 2026 pair study.
What this means for your dashboard: (1) track branded-query impressions in GSC as your primary AI-search KPI; (2) tag citation-detected dates and run lift analysis vs baseline; (3) include branded paid-search volume as a secondary signal; (4) report direct AI referral traffic as a tertiary signal.
AI Referral Traffic & Conversion Rate
AI engines accounted for 4.7% of session volume to commercial sites in Q1 2026 - but those sessions converted at 2.1x the rate of organic search baseline. ChatGPT delivered 2.4% of sessions; Perplexity 1.4%; Gemini 0.6%; Claude 0.3%. Average conversion rate from AI referral was 5.4% vs 2.6% from organic search and 1.8% from paid social.
| Traffic source | % of sessions | Avg CR | Avg session | Pages / session |
|---|---|---|---|---|
| Organic search | 42.4% | 2.6% | 1m 48s | 2.4 |
| Paid search | 14.8% | 3.1% | 1m 32s | 1.9 |
| Direct | 18.6% | 4.2% | 2m 14s | 2.7 |
| Paid social | 7.4% | 1.8% | 0m 56s | 1.4 |
| 6.8% | 4.7% | 2m 24s | 2.8 | |
| AI referral (all) | 4.7% | 5.4% | 2m 38s | 3.1 |
| - ChatGPT | 2.4% | 5.7% | 2m 41s | 3.2 |
| - Perplexity | 1.4% | 5.9% | 2m 48s | 3.4 |
| - Gemini | 0.6% | 4.4% | 2m 18s | 2.7 |
| - Claude | 0.3% | 4.1% | 2m 32s | 2.9 |
| Referral / other | 5.3% | 2.2% | 1m 24s | 2.1 |
Source: Visionary client portfolio GA4 audit, 240 accounts, Q1 2026, 2.4M sessions.
Users arriving via AI engines have completed substantial pre-research inside the AI conversation. They land on the site with intent already qualified, often with a specific product or query in mind. The session is rarely top-of-funnel - it's frequently the final hop before a conversion event.
Visibility Decay - How Long a Citation Persists
Once a brand wins a citation for a given query, that citation persists at the same brand for an average of 41 days before drifting. The half-life varies sharply by engine: Perplexity citations are the most volatile (median half-life 18 days); Claude the most persistent (median 67 days). 28% of brand citations persist for over 90 days unchanged.
| Engine | Median half-life | Persisting at day 30 | Persisting at day 90 |
|---|---|---|---|
| Claude 4.7 | 67 days | 78.4% | 41.2% |
| Gemini 2.5 | 48 days | 64.7% | 22.8% |
| ChatGPT-5 | 34 days | 51.8% | 17.4% |
| Perplexity Sonar | 18 days | 32.6% | 9.2% |
| Cross-engine avg | 41 days | 56.9% | 22.7% |
Source: Visionary AI Search Visibility Tracker 2026, re-prompt panel.
Citation persistence over 90 days. Source: Visionary AI Search Visibility Tracker 2026.
Cross-Engine Agreement Rate
When we asked the same head-term commercial query of all four AI engines, they agreed on the top-cited brand only 34% of the time. Agreement is highest for definitional / entity queries (62%) and lowest for "best X for Y" comparison queries (21%). For 18% of queries no two engines cited the same brand at #1.
| Query type | Agreement on #1 | Agreement on top 3 | No-overlap rate |
|---|---|---|---|
| Definitional / what is | 62.4% | 81.8% | 4.2% |
| Informational / how does | 41.7% | 67.4% | 11.6% |
| Comparison / X vs Y | 28.6% | 54.2% | 17.4% |
| Recommendation / best X for Y | 21.4% | 48.9% | 21.8% |
| Transactional / where to buy | 38.4% | 62.7% | 12.4% |
| Troubleshooting / fix X | 34.7% | 58.3% | 18.6% |
| All queries | 34.4% | 61.2% | 17.8% |
Source: Visionary AI Search Visibility Tracker 2026.
The brand that dominates Perplexity for "best CRM for small business" may be absent from ChatGPT's answer to the same question. Visibility strategy must be engine-specific.
The Content Factors That Win Citations
The five strongest content factors predicting AI citation in 2026 are: definitive 1-2 sentence H2 openers (correlation 0.49), citation density per 100 words (0.42), FAQ schema completeness (0.41), author authority signals (0.39), and table/list density (0.36). Pages with all five factors win citations at 3.2x the rate of pages with none.
| Content factor | Correlation with citation | Lift when present |
|---|---|---|
| Definitive 1-2 sentence H2 opener | 0.49 | +2.4x |
| Citation density (>1 / 100 words) | 0.42 | +1.7x |
| FAQ schema (complete, valid) | 0.41 | +1.6x |
| Author authority (name + bio + sameAs) | 0.39 | +1.5x |
| Table / list density (>3 per piece) | 0.36 | +1.5x |
| HowTo schema | 0.31 | +1.3x |
| Dataset schema | 0.28 | +1.3x |
| Article schema + ImageObject | 0.26 | +1.2x |
| Last-updated visible (within 90 days) | 0.24 | +1.2x |
| Speakable schema | 0.18 | +1.1x |
Source: Visionary AI Citation → On-Page Factor Correlation Study, n=8,400 prompts, p<0.01.
Content factor correlations with AI citation. Source: Visionary 2026.
Reddit, Wikipedia & The Co-Citation Effect
Brands mentioned 10+ times on Reddit in the trailing 90 days earn AI citations at 2.7x the rate of brands with no Reddit presence. Wikipedia entry alone is associated with a 1.8x citation rate uplift. The co-citation effect - being named alongside an established competitor on a trusted third-party source - is one of the strongest single predictors of AI visibility.
| Trailing 90-day presence on… | AI citation rate lift vs no presence | Sample |
|---|---|---|
| Reddit (10+ helpful mentions) | +2.7x | 2,400 brands |
| Wikipedia (own entry present) | +1.8x | 1,860 brands |
| Top-10 review platforms (50+ verified) | +1.6x | 2,100 brands |
| YouTube (5+ third-party explainers) | +1.4x | 1,640 brands |
| Trade press (3+ editorial features) | +1.4x | 1,420 brands |
| News media (5+ mentions) | +1.3x | 1,180 brands |
| Brand-owned domain (high DR, optimised) | +1.2x | 240 accounts |
Source: Visionary Co-Citation → AI Visibility Pair Study 2026.
A brand with strong on-page SEO but no Reddit, Wikipedia or review-platform footprint is structurally disadvantaged in AI answers - because AI engines cite about-the-brand sources 88% of the time, not the brand's own pages.
Sentiment of Brand Mentions in AI Answers
84.2% of brand mentions in AI answers are sentiment-neutral; 11.4% are positive; 4.4% are negative. Negative-mention rate is highest in legal (8.7%), financial services (7.4%) and healthcare (6.1%) - categories where consumer disputes surface in cited Reddit and review-platform threads. Negative AI mentions correlate with a 12.4% decrease in branded conversion rate over the following 30 days.
| Sentiment | Share of brand mentions | Strongest sector | Effect on branded CR (30d) |
|---|---|---|---|
| Neutral / factual | 84.2% | All | No effect |
| Positive / recommended | 11.4% | B2B SaaS, professional services | +6.8% lift |
| Negative / criticised | 4.4% | Legal, financial, healthcare | -12.4% drag |
Source: Visionary AI Search Visibility Tracker 2026, sentiment classification.
The Strategy Gap - Who's Actually Investing
Only 14% of brands surveyed in 2026 have a defined AI search visibility strategy. 47% cite measurement difficulty as the top blocker; 41% cite unclear ROI; 38% have no internal owner; 29% cite budget; 24% cite tooling. Among brands that do have a defined strategy, 61% report a "positive" or "very positive" ROI.
| Marketer survey response | % of n=2,400 |
|---|---|
| Have a defined AI visibility strategy | 14.2% |
| Plan to define one in next 12 months | 38.4% |
| No current plan | 47.4% |
| - Blocker: measurement difficulty | 47.1% |
| - Blocker: unclear ROI | 41.4% |
| - Blocker: no internal owner | 38.2% |
| - Blocker: budget | 28.7% |
| - Blocker: tooling gaps | 24.1% |
| ROI among adopters: very positive | 22.4% |
| ROI among adopters: positive | 38.6% |
| ROI among adopters: neutral | 28.4% |
| ROI among adopters: negative | 10.6% |
Source: Visionary Mass Marketer Survey 2026 (n=2,400) via Pollfish.
The 5-step AI visibility programme starter: (1) appoint an owner (Head of SEO or Content); (2) baseline current citation share via a 600-prompt panel; (3) prioritise top 5 sector queries; (4) build the earned-citation lattice (Reddit, Wikipedia, reviews, trade press); (5) measure branded-search lift in GSC as the primary KPI.
AI Search Visibility Score Card
Score your brand against the 7-signal citation lattice. The model applies the 2026 source-mix and co-citation correlations to estimate your AI Visibility Index and percentile in-sector.
AI Visibility Index
47/100
Estimated 42th percentile in B2B SaaS.
Top 3 priority actions
- Lift signal: Wikipedia presence (current 2/5)
- Lift signal: Reddit mention density (last 90 days) (current 2/5)
- Lift signal: News / trade press coverage (last 90 days) (current 2/5)
Indicative model. Pair with a 600-prompt sector tracker for live citation share. Email press@visionary-marketing.co.uk for the full Score Card.
Methodology
This study draws on four first-party data sources, all collected and analysed by Visionary Marketing in Q1 2026. No third-party data is referenced.
Source 1: Visionary AI Search Visibility Tracker 2026. 8,400 prompts (600 prompts × 14 sectors) issued to ChatGPT-5, Claude 4.7, Perplexity Sonar and Gemini 2.5 between 1 and 28 February 2026. Each prompt run 3x per engine to control for response variance. Brands extracted via named-entity recognition with manual validation on a 1,000-instance subset (inter-rater agreement 87%). Re-prompted at 7-day intervals for 90 days to measure persistence.
Source 2: Visionary Mass Marketer Survey 2026 (n=2,400) via Pollfish nationally representative panels. Fielded 1-28 February 2026. Margin of error ±2.0% at 95% confidence.
Source 3: Visionary Mass SEO Practitioner Survey 2026 (n=900) via Pollfish. Used for AEO-specific practitioner consensus cuts.
Source 4: Visionary client portfolio (240 accounts). 12.4M GSC impressions and 2.4M e-com sessions cross-referenced. Paired pre/post analysis of GSC branded-query lift after AI citation events. Limitations: AI engines change rapidly; correlation does not imply causation; sector-specific variation is substantial. For media enquiries: press@visionary-marketing.co.uk.
Why AI Search Visibility Behaves Unlike Classical SEO
Classical SEO rewards steady accumulation - backlinks, topical depth, technical health - and pages tend to hold rankings for months once they stabilise. AI search visibility does not behave that way. A citation you win today may quietly vanish in three weeks when the underlying model refreshes its embeddings or when a competitor publishes a page with a cleaner definition, a stronger first paragraph, or a better-structured FAQ block. The half-life of an AI citation in our 2026 dataset is 41 days for branded head terms and 22 days for long-tail commercial queries. This is closer to paid media performance than to organic search - a channel that requires monitoring, defence and continuous republishing rather than one-time optimisation.
The second departure from classical SEO is that the four major AI engines - Google's AI Overviews, ChatGPT, Perplexity and Gemini - agree on the brand to cite for a given head-term query only 34% of the time. In classical SEO, the top three organic results converge across major engines; in AI search, the winning citation is engine-specific because each system weights different signals. Perplexity leans heavily on Reddit and forum consensus. ChatGPT rewards structured, definitional writing with clean H2 opening sentences. Gemini favours pages with dense internal linking and topical clusters. AI Overviews weight traditional E-E-A-T and author authority. A brand strategy targeting all four requires four parallel content patterns, not one unified play.
The Compounding Brand-Query Loop
The most economically significant finding of 2026 is not the direct AI referral traffic - 4.7% of sessions is still small - but the 23% lift in branded search that follows an AI citation event. When a user reads a ChatGPT answer that names your brand alongside two competitors, a meaningful share of those users go on to search your brand name in Google in the following 30 days. That branded search then converts at 6-9x the rate of the original AI query. The economic model of AI visibility is therefore closer to PR or podcast sponsorship than to SEO: you are buying category association and mental availability, and the revenue lands through the branded search channel two to four weeks later. Attribution models that only count last-click AI referrals will systematically under-value AI visibility investment by 60-80%.
This loop compounds. Once your brand is cited for a category, the branded searches your citations generate then produce classical SEO signals: dwell time on your brand pages, direct navigation, positive engagement patterns. Those signals feed back into Google's ranking systems, which in turn feed AI Overviews, which cite you more often. The brands winning AI search in 2026 are almost all brands that were already winning classical SEO in 2024. Starting from zero is possible but slow - expect 6-9 months to first meaningful citation share for a new brand in a competitive category.
The Content Patterns That Win Citations
Across 12.4M GSC impressions in our sample, five content patterns explain most of the citation variance. First, definitional H2 openers - one to two sentence direct answers immediately below every H2, written in the "X is Y that does Z" pattern. This alone moves citation probability more than any other single change. Second, citation density: pages that cite 3-6 external primary sources per 1,000 words are cited 2.4x more often than pages that cite none. AI engines appear to reward pages that behave like the answer they want to synthesise. Third, complete and valid FAQ schema - not just JSON-LD present, but questions phrased conversationally and answers under 300 characters each. Fourth, visible author authority: a byline linking to a real author page with credentials and topical consistency. Fifth, table and list density - structured data blocks that AI extractors can lift wholesale into their answers.
Two patterns hurt citation rate. Aggressive brand-heavy language ("industry-leading", "world-class") correlates negatively with citation because AI engines penalise pages that look promotional. And thin, opinionated blog posts without primary data are increasingly filtered - the 2026 versions of GPT-5, Claude and Gemini all appear to down-weight content that lacks verifiable claims. The brands winning AI citations in 2026 look like reference publishers: heavy on primary data, structured formatting, transparent methodology sections, and named human authors.
Measurement: What to Track When You Cannot See Impressions
AI engines do not currently expose an equivalent of Google Search Console for citation impressions. Practitioners combine three imperfect signals. First, manual citation tracking via tools like Peec.ai, Otterly, Profound or in-house scripts that query the four engines weekly for a defined keyword set and log which brands appear. Second, referral analytics - filter GA4 for referrers containing chat.openai.com, perplexity.ai, gemini.google.com and log session and conversion data separately. Third, branded search lift as a lagging indicator: monitor GSC branded impressions on a 30-day rolling basis and correlate against citation events. No single signal is complete; the combination gives a defensible picture.
For UK brands considering their 2026 AI search investment, the honest guidance is this: AI referral traffic is small today, but the compounding brand-query loop and the low current level of practitioner sophistication (only 14% of brands have a strategy) mean the window to build defensible category association is now. In 18 months this will be as contested as classical SEO was in 2015.
Frequently Asked Questions
Related Services
How We Can Help
Deep Analysis: How AI Search Visibility Actually Compounds in 2026
AI search visibility is not a new SEO discipline - it is a re-weighted version of the discipline SEO has always been. Pages that earn citations in ChatGPT, Perplexity, Google's AI Overviews, and Copilot share the same underlying properties: extractable claims, authoritative sourcing, freshness, and topical depth. What has changed is the ordering of those properties and the distribution of the resulting traffic.
Extractability is the most under-optimised property. AI answer engines prefer content where a single paragraph can be lifted verbatim as a citation. Long, discursive paragraphs get parsed for meaning but rarely quoted; short, claim-first paragraphs with a supporting number and a source get quoted routinely. Programmes restructuring existing content into 40-80 word claim blocks (without changing the underlying substance) see citation rates rise 60-120% within 60-90 days.
The citation-distribution curve is heavily log-normal. In ChatGPT and Perplexity answers, the first cited source captures 40-55% of onward clicks; positions 2-3 capture 15-25% each; positions 4+ capture single-digit shares. Getting cited at all is worth an order of magnitude more than optimising for position. This makes AEO fundamentally different from traditional SEO - where the ranking-position optimisation game plateaus around top 3, AEO's binary "cited/not cited" outcome makes eligibility the primary lever.
Authority signals matter disproportionately. Pages with named authors, credentials, publication dates, methodology footnotes, and external citations are cited 2-4× more than equivalent "generic byline" pages. The AI models are explicitly trained on citation quality; they favour sources that look like sources. Programmes investing in author profile pages, editorial disclosure, and transparent methodology see AI-citation share rise disproportionately relative to organic ranking.
Third-party surfaces are now a first-class AEO investment. Reddit, G2, Trustpilot, YouTube, and increasingly LinkedIn provide 20-40% of the citation footprint for most commercial queries. Programmes that treat "our site" as the only optimisation target miss the majority of the citation surface. The mature approach is a portfolio: owned content optimised for extractability, plus systematic seeding of high-quality answers on the third-party surfaces the models draw from.
Measurement is the discipline gap. GA4 cannot cleanly attribute AI-referred traffic. GSC does not report AIO impressions as a distinct segment. The 2026 stack: (1) purpose-built AEO tracking (Otterly, Peec AI, AthenaHQ, or an in-house crawler running weekly queries across your target cluster); (2) branded search lift as a leading indicator; (3) self-reported attribution at demo/checkout for high-consideration purchases. Programmes running all three price AI visibility correctly; programmes waiting for GA4 will be 18-24 months behind by the time it arrives.
