{
  "schema": "demandbox.citations/1",
  "generatedAt": "2026-08-28T15:51:47.690Z",
  "piece": {
    "kind": "guide",
    "slug": "what-is-answer-engine-optimization",
    "title": "What Is Answer Engine Optimization (AEO)? The Complete 2026 Guide",
    "url": "https://demandbox.co/guides/what-is-answer-engine-optimization",
    "published": "2026-08-01",
    "updated": "2026-08-05",
    "lastReviewed": "2026-08-05"
  },
  "claims": [
    {
      "blockIndex": 9,
      "blockType": "stat",
      "value": "5 criteria",
      "claim": "A documented re-ranking pass sits between search retrieval and citation selection, applying source type credibility, consensus detection, evaluative depth, a se…",
      "sourceId": "sigiRerank",
      "sourced": true,
      "hedged": false,
      "confidence": "sourced"
    },
    {
      "blockIndex": 10,
      "blockType": "stat",
      "value": "21,143 citations",
      "claim": "A measurement framework across 602 prompts found substantial gaps between what search layer retrieval surfaces and what the generated answer actually cites.",
      "sourceId": "geoCitationAbsorption",
      "sourced": true,
      "hedged": false,
      "confidence": "sourced"
    },
    {
      "blockIndex": 14,
      "blockType": "stat",
      "value": "252,000 trials",
      "claim": "Controlled paired retrieval trials isolating 18 content factors found structural and claim level properties, not overall page quality, drove selection between t…",
      "sourceId": "geoWhatGetsCited",
      "sourced": true,
      "hedged": false,
      "confidence": "sourced"
    },
    {
      "blockIndex": 17,
      "blockType": "stat",
      "value": "Rank conditioned",
      "claim": "An audit of Google AI Overviews on high stakes queries found citation selection is conditioned on both search rank and source provenance, so ranking is an advan…",
      "sourceId": "aiOverviewsAudit",
      "sourced": true,
      "hedged": false,
      "confidence": "sourced"
    },
    {
      "blockIndex": 21,
      "blockType": "stat",
      "value": "25m sessions",
      "claim": "Analysis of 25 million B2B web sessions between September 2025 and March 2026 measured AI assistant referral traffic growing as a share of research entry points…",
      "sourceId": "octane11",
      "sourced": true,
      "hedged": false,
      "confidence": "sourced"
    },
    {
      "blockIndex": 22,
      "blockType": "stat",
      "value": "8 months",
      "claim": "Across 53 B2B SaaS brands, AI referral traffic grew while total organic sessions flattened, meaning session counts alone understated the channel's contribution …",
      "sourceId": "piperocketSaas",
      "sourced": true,
      "hedged": false,
      "confidence": "sourced"
    },
    {
      "blockIndex": 44,
      "blockType": "stat",
      "value": "25,337 citations",
      "claim": "Across 21,075 AI engine responses in eight industries, third party sources including review platforms and community threads were cited far more often than vendo…",
      "sourceId": "deltavCitations",
      "sourced": true,
      "hedged": false,
      "confidence": "sourced"
    },
    {
      "blockIndex": 47,
      "blockType": "stat",
      "value": "Feature level",
      "claim": "Optimising structural, content, and linguistic features outperformed token level rewriting for generative citation visibility, which is why format changes beat …",
      "sourceId": "featGeo",
      "sourced": true,
      "hedged": false,
      "confidence": "sourced"
    },
    {
      "blockIndex": 55,
      "blockType": "table",
      "value": "1 to 3 weeks / 2 to 6 weeks / 2 to 6 weeks / 1 to 2 quarters",
      "claim": "Work type | First signal | Durable effect",
      "sourced": false,
      "hedged": false,
      "confidence": "illustrative-model"
    }
  ],
  "sources": [
    {
      "id": "geoWhatGetsCited",
      "title": "What Gets Cited: Competitive GEO in AI Answer Engines",
      "publisher": "arXiv",
      "published": "2026-07-15",
      "kind": "study",
      "method": "252,000 paired retrieval trials across six large language models, isolating 18 content factors one at a time in a two document RAG testbed.",
      "url": "https://arxiv.org/html/2605.25517",
      "finalUrl": "https://arxiv.org/html/2605.25517",
      "linkStatus": "ok",
      "linkCheckedAt": "2026-08-26T14:52:46.360Z",
      "freshness": "fresh",
      "freshnessWindowMonths": 18,
      "expiresOn": "2028-01-15"
    },
    {
      "id": "geoCitationAbsorption",
      "title": "From Citation Selection to Citation Absorption: A Measurement Framework for Generative Engine Optimization",
      "publisher": "arXiv",
      "published": "2026-04-28",
      "kind": "study",
      "method": "602 prompts, 21,143 search layer citations, and 23,745 citation level feature records across major AI search platforms.",
      "url": "https://arxiv.org/html/2604.25707",
      "finalUrl": "https://arxiv.org/html/2604.25707",
      "linkStatus": "ok",
      "linkCheckedAt": "2026-08-26T14:52:46.360Z",
      "freshness": "fresh",
      "freshnessWindowMonths": 18,
      "expiresOn": "2027-10-28"
    },
    {
      "id": "sigiRerank",
      "title": "Search Position Versus Citation Priority: Evidence for a Separate Re-Ranking Pass in AI Answer Generation",
      "publisher": "Scientific Institute for Generative Intelligence (SIGI-2026-056)",
      "published": "2026-03-01",
      "kind": "study",
      "method": "Observational study documenting five re-ranking criteria applied between retrieval and citation: source type credibility, consensus detection, evaluative depth, self ranking discount, and claim specificity.",
      "url": "https://generativeintelligence.institute/publications/SIGI-2026-056/",
      "finalUrl": "https://generativeintelligence.institute/publications/SIGI-2026-056/",
      "linkStatus": "ok",
      "linkCheckedAt": "2026-08-26T14:52:46.360Z",
      "freshness": "fresh",
      "freshnessWindowMonths": 18,
      "expiresOn": "2027-09-01"
    },
    {
      "id": "aiOverviewsAudit",
      "title": "Auditing Citation Behavior in AI-Generated Search Summaries: A Case Study of Google AI Overviews",
      "publisher": "Canadian Conference on Artificial Intelligence (PMLR 318)",
      "published": "2026-06-29",
      "kind": "study",
      "method": "Rank and provenance conditioned analysis of Google AI Overviews citations on high stakes queries drawn from MS MARCO Web Search.",
      "url": "https://proceedings.mlr.press/v318/kakimov26a.html",
      "finalUrl": "https://proceedings.mlr.press/v318/kakimov26a.html",
      "linkStatus": "ok",
      "linkCheckedAt": "2026-08-26T14:52:46.360Z",
      "freshness": "fresh",
      "freshnessWindowMonths": 18,
      "expiresOn": "2027-12-29"
    },
    {
      "id": "featGeo",
      "title": "Think Before Writing: Feature-Level Multi-Objective Optimization for Generative Citation Visibility",
      "publisher": "ACL 2026 (Long Papers)",
      "published": "2026-07-02",
      "kind": "study",
      "method": "Feature level optimization of structural, content, and linguistic page properties, compared against token level rewriting for citation visibility.",
      "url": "https://aclanthology.org/2026.acl-long.929.pdf",
      "finalUrl": "https://aclanthology.org/2026.acl-long.929.pdf",
      "linkStatus": "ok",
      "linkCheckedAt": "2026-08-26T14:52:46.360Z",
      "freshness": "fresh",
      "freshnessWindowMonths": 18,
      "expiresOn": "2028-01-02"
    },
    {
      "id": "deltavCitations",
      "title": "AI Search Citations Study: What 25,000+ Citations Reveal",
      "publisher": "DeltaV Digital",
      "published": "2026-07-13",
      "kind": "ai-search",
      "method": "21,075 AI engine responses and 25,337 citations tracked across ChatGPT, Perplexity, Gemini, Google AI Overviews, and AI Mode in eight industries between 14 April and 13 July 2026.",
      "url": "https://www.deltavdigital.com/resources/reports/ai-citation-study/",
      "finalUrl": "https://www.deltavdigital.com/resources/reports/ai-citation-study/",
      "linkStatus": "ok",
      "linkCheckedAt": "2026-08-26T14:52:46.360Z",
      "freshness": "fresh",
      "freshnessWindowMonths": 3,
      "expiresOn": "2026-10-13"
    },
    {
      "id": "octane11",
      "title": "State of B2B AI Search, Vol. 1",
      "publisher": "Octane11",
      "published": "2026-04-08",
      "kind": "ai-search",
      "method": "25 million B2B web sessions analysed between September 2025 and March 2026, measuring referral traffic by AI engine.",
      "url": "https://www.octane11.com/insights/insights-state-of-b2b-ai-search/",
      "finalUrl": "https://www.octane11.com/insights/insights-state-of-b2b-ai-search/",
      "linkStatus": "ok",
      "linkCheckedAt": "2026-08-26T14:52:46.360Z",
      "freshness": "stale",
      "freshnessWindowMonths": 3,
      "expiresOn": "2026-07-08"
    },
    {
      "id": "piperocketSaas",
      "title": "AI SEO Statistics: B2B SaaS Traffic and Lead Data",
      "publisher": "PipeRocket Digital",
      "published": "2026-07-01",
      "kind": "ai-search",
      "method": "Analytics and CRM data from 53 B2B SaaS brands tracked over eight months, comparing organic search against AI referral traffic on traffic, leads, and pipeline.",
      "url": "https://piperocket.digital/research/ai-seo-statistics/",
      "finalUrl": "https://piperocket.digital/research/ai-seo-statistics/",
      "linkStatus": "ok",
      "linkCheckedAt": "2026-08-26T14:52:46.360Z",
      "freshness": "fresh",
      "freshnessWindowMonths": 3,
      "expiresOn": "2026-10-01"
    }
  ],
  "revisions": [
    {
      "date": "2026-08-05",
      "summary": "Expanded to the canonical explainer: the four step citation pipeline, engine by engine behaviour, what changed in 2026, the full workflow, a starter checklist, and links to every related guide."
    },
    {
      "date": "2026-08-01",
      "summary": "First published."
    }
  ],
  "totals": {
    "claims": 9,
    "sourced": 8,
    "unsourced": 1,
    "staleSources": 1
  }
}