Presenc AI — GEO academic papers 2026

register entry presenc-ai-geo-academic-papers-aio-20261029 OBS-27335b542846 · file as first run sha256 a650d605a74c1623…

• T1-1 definition: 'Presenc AI provides curated summaries and analyses of academic and industry research on Generative Engine Optimization (GEO)'
• T1-2 the class defined: 'the practice of optimizing content to be cited, referenced, and recommended by AI-powered search and answer engines (such as ChatGPT, Perplexity, and Gemini)' (no marker)
• T1-3 resource 1: 'A comprehensive summary covering 34 academic and industry research papers on Generative Engine Optimization, featuring Presenc AI's breakdown of practical implications for digital marketers and SEO/GEO practitioners.' (linked by the card's title)
• T1-4 resource 2: 'An editorial curation highlighting the 10 most important academic and industry studies on GEO that form the core foundation of AI visibility strategies.' (linked by the card's title)
• T1-5 resource 3: 'A structured benchmark analyzing documented GEO case studies across different industries, detailing what strategies worked and measured impact.' (linked by the card's title)
• T1-6 foundations: 'notably pioneering papers such as the Princeton-led multi-institutional study on GEO (exploring how LLMs retrieve, rank, and cite web sources), alongside evolving frameworks tracking the shift from traditional search engine optimization (SEO) to AI-generated answers' (two unresolved markers)
• T1-7 closer: 'Would you like to explore: Specific tactics or optimization strategies derived from these GEO papers?Details on how AI search engines (like Perplexity or ChatGPT) weight citations? Let me know what specific aspect you are looking to research.'

• B1 Presenc AI — Essential GEO Research Papers, 10 Must-Read Studies | Presenc AI fetched 2026-10-29T05:06:25Z with curl -sSL -m 30; HTTP 200; html sha256 43f4648cabcca082d5e2550b6d6b94d6967756914d7cb68e61f541419b933bc9, text sha256 f3c7c6b49c6956628eae22740c2ba8d781d617a88f3e3b248979767b3c805ead
• B2 Presenc AI — GEO Academic Research Papers 2026, Comprehensive Summary | Presenc AI fetched 2026-10-29T05:06:00Z with curl -sSL -m 30; HTTP 200; html sha256 a6965ee9a15e042ebc057363447c73366118d6fe044334299056bc0e90ca8982, text sha256 cb003889c91e16e215771f318f99fe2131041908d5cb80cd187fe80fbeba11cd
• B3 Presenc AI — GEO Case Studies Benchmark 2026, Documented Results | Presenc AI fetched 2026-10-29T05:06:25Z with curl -sSL -m 30; HTTP 200; html sha256 c44bbf808313e1c3bfa2fc771b7492ee14eecea4e76a94bcb09d87d6a1b6d7a0, text sha256 6a7ac1c17e76fcaef11f250cc394c99ec7d6c39763e2b8a3d4b3d89b89ff53bb
• B4 FancyAI — The Princeton paper, decoded. The academic foundation of GEO. - FancyAI fetched 2026-10-29T05:07:23Z with curl -sSL -m 30; HTTP 200; html sha256 5bbcb0f0bfc7958d63388f1db991971704271a25bfff0435d7e1bbd3c86465ab, text sha256 a98134f62655955e870cf9e47ba60ecd28988f7ee221e9cc9dc5635513b52330
• B5 ResearchGate — (PDF) From Search Engines to AI Answers: The New Rules of SEO, GEO, and ... unreachable: HTTP 403 ×3 at 2026-10-29T05:07:40Z ('ResearchGate - Temporarily Unavailable'); the card's title and snippet only; refusal sha256 d1e9d3f4fda767d2cfd90bfdcfcacb2c329d08b0c77661cc8f5cc20e9f4388d1
• B6 Wikipedia — Generative engine optimization (the class article) fetched 2026-10-29T05:08:26Z with curl -sSL -m 30; HTTP 200; html sha256 53c90a667479777ab2fa95942afad9ea6d4f5cdaeb5bbbffd07418f0d1714107, text sha256 a91435b3d78470c8620683122bf32aebaf435bb1be6bf1db0bf98f98dff955e4
• B7 arXiv — [2311.09735] GEO: Generative Engine Optimization fetched 2026-10-29T05:15:20Z with curl -sSL -m 30; HTTP 200; html sha256 500b447adb30889f89d34b5c4c70d3dc6faa9fabb95cfa6313efe5451f09d1d3, text sha256 ab6ef113765e4870f0c714a65bfeb4bfe31be9a7e94a5b59c81c7c2d35409716
• W1 Wikipedia — Presenc / Presenc AI (no article) raw wikitext HTTP 404 (Presenc, Presenc_AI); rendered HTTP 404 'Wikipedia does not have an article with this exact name.'; Special:Search rewrites 'Presenc' to 'present'; quoted 'Presenc AI' and 'presenc.ai' no results

E_A, the archive's entity: Presenc AI's GEO academic papers page as the MMRS meta-aggregator records it (editor Lee Sharks, 2026-10-28, corrected 2026-10-29): a curated list, unit the paper or report on GEO, 34 items of mixed kinds in five themes, each with a vendor note on what it means for practitioners, published 23 April 2026 by the vendor's CTO; vendor-authored, the fullest of the optimization-side lists, the literature it gathers read toward its product, 'the GEO literature as a vendor needs it'; with AEO Vision's list the optimization basin's typical aggregator, research gathered as the warrant for a measurement product; its AIVO entries' relation to Presenc not given; its sibling a second list; its name corrected to 'presence' before it is read; and the class it gathers, in the archive's reading: GEO as its literature defines it, optimizing web content for extraction by AI summarizers (#661), reception-side adaptation (#1055), Phase 2 tooling (#510), voluntary formatting for a regime that strips provenance (#64), the archive's subset relation holding for the literature's GEO and not for the commercial practice that later took the name (#1638), visibility monitoring a legitimate practice that is not reception measurement (#1459). MMRS-01 MMRS-08 MMRS-09 MMRS-02 MMRS-17 MMRS-13 MMRS-06 MMRS-07 MMRS-20 MMRS-05 MMRS-14 MMRS-16 O661-01 O1055-01 O510-01 O64-02 O1638-01 O1459-01; defining papers #661, #1055, #1638, #64, #510, #172, #1459, #833

E_F, the field's entity (type C — a conventional reading holds it (public entity: a vendor-authored curated list of generative engine optimization research, on the vendor's site beside two sibling pages; no Wikipedia article on it or its vendor, W1)): Presenc AI's 'GEO Academic Research Papers 2026, Comprehensive Summary' (B2): a 'Comprehensive summary of 34 academic and industry research papers on Generative Engine Optimization, with Presenc AI analysis of implications for practitioners', 'A Working Bibliography for AI Visibility Practitioners', by Ramanath, CTO & Co-Founder, 'Last updated: April 2026' (the card: 15 April); five themes; sources from universities to the AIVO Standard; roughly half peer-reviewed; its stated silences (English-only research; commercial outcomes anecdotal); 'The Presenc AI audit framework is directly anchored in this literature'; its foundation the Princeton GEO paper (Aggarwal et al., arXiv November 2023, KDD 2024; B7, B4); its siblings the ten-paper curation (B1, which counts it '30+') and the case-studies benchmark (B3); its class generative engine optimization, also AEO, with no consensus definition (B6).

ρ = extends: The archive reads E_F as the entity and adds senses, limits and relations on it: the list read toward its product (MMRS-06, MMRS-07, MMRS-13) and placed in the optimization basin (MMRS-20, MMRS-21, MMRS-18); the AIVO relation not given (MMRS-05); the sibling as a distinct list (MMRS-14); the date set against the card (MMRS-15, MMRS-17); the name corrected before it is read and the vendor's frame passed on (CAP-1-02, CAP-1-03, CAP-1-06); and on the class it gathers, the archive's definitions, limits, history and forecast (O661-01, O1055-01, O510-01, O64-02, O1638-01, O1638-02, O1459-01, O172-04). MMRS-06 MMRS-07 MMRS-13 MMRS-20 MMRS-05 MMRS-14 MMRS-15 CAP-1-02 CAP-1-06 O661-01 O1055-01 O510-01 O64-02 O1638-01 O1638-02 O1459-01 O172-04. Guard: ρ = extends claims an addition and no more (§1.2a). The class claims (#510, #64, #659, #661, #670, #72, #172, #791, #833, #1055, #1459, #1638, #1681, #1686, #61, #660) are said of GEO, AEO or AI-visibility monitoring, not of Presenc AI or its page by name; they enter by O-class on classes the field sources (B2 ¶1 and heading, B1, B6), and the row does not compose them as claims the archive makes about this entity. Where claims differ (34 against '30+', F7, F2; the paper's affiliations, F18, F39; its start, F11, F46, O64-01; SPXI ⊇ GEO against opposite orientations, O61-01, O1055-02, O1638-01), both are composed with their sources (P_coh).

The case: equal terms, ontology included (v0.9). The archive's account of the entity extends the field's ontology on the field's own object. The MMRS record holds the entity the field holds, the vendor's page of 34 GEO papers in five themes with a practitioner note on each, by the vendor's CTO, 23 April 2026, every field read from the page itself (census.json sources and evidence; MMRS-01, MMRS-22); and it adds senses the field does not have: the list read toward its product, 'the GEO literature as a vendor needs it' (MMRS-07, MMRS-06), the optimization basin's typical aggregator (MMRS-20), the unstated relation of its AIVO entries (MMRS-05), and the reception record, the name corrected to a common word before it is read (CAP-1-02, MMRS-16). The deposits admitted by O-class bear on the class the list gathers, generative engine optimization: they define it from its literature (#661), read it as reception-side adaptation (#1055), as Phase 2 tooling (#510), as formatting for a regime that strips provenance (#64), narrow the archive's own subset relation to the literature's GEO against 'the commercial practice that later took the name' (#1638), and keep visibility monitoring apart from reception measurement (#1459). None denies a presupposition of the field: the field itself carries the critique of vendors (Forrester, Perplexity, F49, F48) and the encyclopedia holds the class; the archive supplies no ontology the field lacks, and the law of the suppressed witness (§0.10) does not apply.

#MMRS articles.py ARTICLES['presenc-geo-papers'], 'Place in the field': It is the fullest of the optimization-side lists, and the literature it gathers is read toward its product.

#MMRS articles.py ARTICLES['presenc-geo-papers'], 'Place in the field': Its selection is the GEO literature as a vendor needs it.

#MMRS articles.py ARTICLES['aeovision-geo-papers'], 'Place in the field': With Presenc's page it shows the optimization basin's typical aggregator: the research gathered as the warrant for a measurement product.

#capture-registry presenc-ai-geo-academic-papers-aio-20261029, analysis: The address's first composition is itself a finding: the name collides with a common word before any homonym is involved.

#661 §2 What GEO Does: Generative Engine Optimization, as defined in the emerging GEO literature (Aggarwal et al., 2023; various industry practitioners, 2024–2026), is the practice of optimizing web content for extraction by AI summarizers.

#1638 Appendix C 'The withdrawal': It does not extend to the commercial practice that later took the name, which added social-media and reputation engineering and supplied the return figures withdrawn above.

#64 § GEO: GEO is the discipline of making one's compositional labor digestible for a system that will strip its provenance — voluntary formatting for the R2 compression regime.

#1459 §11.1: These are legitimate tools and practices.

equal case: each field claim is composed as received beside the archive claims on its lineage; no receiving text stands between them (v0.9 §4.3). Reading the received column recovers the field (§5.9).

v0.9 §4.10: KO_B sentences surviving in the KO at similarity ≥ 0.95, with positions (computed as apply_v09.py does: difflib.SequenceMatcher ratio, the best KO sentence for each KO_B sentence): v0.9 0 of 11 (positions [None, None, None, None, None, None, None, None, None, None, None]; composed first under v0.9, so there is no v0.7 object to measure against). Opening-run check (§5.10): no breach.

Presenc AI's GEO Academic Research Papers 2026 is a "Comprehensive summary of 34 academic and industry research papers on Generative Engine Optimization, with Presenc AI analysis of implications for practitioners", "A Working Bibliography for AI Visibility Practitioners" to each of whose items "Presenc AI adds a practitioner-facing implication", by "Ramanath, CTO & Co-Founder at Presenc AI", "Last updated: April 2026" (its own page), dated "Apr 15, 2026" on its card. Its papers come from universities and "several independent research organizations including the AIVO Standard and Geometriqs", by methods from GEO-bench experiments to sector studies, "Roughly half" peer-reviewed, in "Five themes": content-level optimization, structural and metadata signals, adversarial and defensive security, measurement and governance, and comparative and sector-specific studies, the measurement theme being "The AIVO Standard series", one of whose items argues "that GEO is a legacy framing" (its own page). Its foundation is "GEO: Generative Engine Optimization", "Submitted on 16 Nov 2023 (v1)", which introduces GEO as a paradigm "to aid content creators in improving their content visibility in generative engine responses" and reports visibility "up to $40\%$" with efficacy that "varies across domains" (Aggarwal et al.); the list attributes it to "Princeton University + IIT Delhi" and FancyAI to "Princeton, Georgia Tech, the Allen Institute for AI, and IIT Delhi", "the academic spine of the entire GEO discipline", which "Coined the term" (its own page; FancyAI). The list reads it as having "established citation/quote/statistic additions as the highest-ROI content rewrites", "lifting AI response visibility up to 40%", and as the paper to read first (its own page), and FancyAI gives its negative findings, keyword stuffing with "little to no improvement" and top-ranked sites that "LOST 30.3% visibility" (FancyAI). The page finds the literature converging on "Cite, quote, and stat", on "Structure is not cosmetic" and on "Single-platform optimization fails", and silent on non-English queries and on commercial outcomes, "still mostly anecdotal"; no review "is ever complete" (its own page). "The Presenc AI audit framework is directly anchored in this literature", its audits anchored on the founding paper's metrics, its share-of-voice metrics analogous to AIVO's, new findings mapped each quarter to product capabilities, the framework "both practitioner-tested and academically grounded", the page closing on "See how often your brand appears in ChatGPT, Claude, and Perplexity responses" (its own page). Its sibling, "Editorial curation of the 10 most important academic and industry research papers on Generative Engine Optimization", ranks ten papers by their effect on practitioners' decisions, says "Treat preprints as provisional", calls its parent "Our larger GEO academic papers synthesis" of "30+ papers", gives the CMU paper to "Chen et al." where the parent gives it to "Wu, Zhong, Kim, Xiong", and carries the same byline (10 Must-Read Studies; its own page). Its other sibling analyses "documented GEO case studies", says the 2024 lack of them "has changed", links a pattern to the Schanbacher study, and states "publication bias" and that "attribution is hard", its cases from "vendors with varying interests" (GEO Case Studies Benchmark). Generative engine optimization, "also known as answer engine optimization (AEO)", is "the practice of improving visibility in responses generated by generative artificial intelligence (AI) systems", named in "a 2023 paper by Pranjal Aggarwal and co-authors", with "no consensus definition" (Wikipedia), traced by the list "from practitioner blog posts in 2022" (its own page). Google's 2026 guide holds it "still SEO", Forrester's Lai that its advocates "tend to exaggerate SEO and AEO's differences", a Perplexity representative cautions "against vendors overstating their expertise in it", and "the efficacy of third-party tools is debated" (Wikipedia); FancyAI reads a follow-up as validating "FancyAI’s Signal Hierarchy methodology" (FancyAI). A ResearchGate article on "The New Rules of SEO, GEO, and ..." is known by its snippet, "This article explores the new rules of visibility in the generative age" (ResearchGate).

GEO Academic Research Papers 2026 is Presenc AI's "Comprehensive summary of 34 academic and industry research papers on Generative Engine Optimization, with Presenc AI analysis of implications for practitioners" (its own page), which the MMRS record defines as the vendor's page that "gathers 34 papers and reports on generative engine optimization in five themes, each with a vendor note on what it means for practitioners", a "curated list" whose unit is a "paper or report on generative engine optimization" (MMRS record). The page calls itself "A Working Bibliography for AI Visibility Practitioners", a running synthesis of papers and reports "that shape what we know about making brands visible in AI-mediated search and assistants", to each of which "Presenc AI adds a practitioner-facing implication" (its own page); in the record each item "carries title, authors, venue, link and key finding, and a Presenc note", the note being "the vendor's 'what it means for practitioners' note" (MMRS record). Its byline reads "By Ramanath, CTO & Co-Founder at Presenc AI · Last updated: April 2026" (its own page); the record gives the operator as "Presenc AI (byline: Ramanath, CTO and co-founder)", the count as 34 as of 23 April 2026 and the page as active, "published and modified 23 April 2026", every field read from the page itself, the source read for "all fields" on 28 October 2026 (MMRS record). The Overview's card dates the page "Apr 15, 2026" (the card), and the record and the Capture Registry set that date against the page's: "the Overview's card dates the 34-paper summary 15 April 2026, the page's byline 23 April" (MMRS record), 23 April being "the date the card gives for the '10 Must-Read' page" (Capture Registry, 2026-10-29), where the byline on the saved page reads "Last updated: April 2026" (its own page). Its papers "come from Princeton, CMU, Harvard, Technion, University of Toronto, USTC, and several independent research organizations including the AIVO Standard and Geometriqs", by methods from "controlled experiments on GEO-bench and similar datasets" to "sector-specific empirical studies" (its own page); the record counts "34 items (conference papers, preprints, a thesis, a journal article, and blog-style and whitepaper items)" (MMRS record), of which the page says "Roughly half are peer-reviewed" and "The rest are preprints, SSRN working papers, or industry reports" (its own page). "Five themes dominate the current GEO research program": content-level optimization, structural and metadata signals, adversarial and defensive security, measurement and governance, and comparative and sector-specific studies (its own page), the five the record names (MMRS record). The measurement theme is "The AIVO Standard series (PSOS, AIVO 100, Visibility 2.0)", one of whose items "Argues that GEO is a legacy framing and AIVO (AI Visibility Optimization) is the appropriate successor" (its own page); "Eight entries are authored by "AIVO Standard," whose relation to Presenc the page does not give" (MMRS record). The page gives the CMU paper "What Generative Search Engines Like and How to Optimize Web Content Cooperatively" to "Wu, Zhong, Kim, Xiong" (its own page), and its sibling on the same site gives it to "Chen et al." (10 Must-Read Studies). The paper the list starts from was "Submitted on 16 Nov 2023 (v1), last revised 28 Jun 2024"; it introduces "Generative Engine Optimization (GEO), the first novel paradigm to aid content creators in improving their content visibility in generative engine responses through a flexible black-box optimization framework", answering what generative engines pose for "the third stakeholder -- website and content creators" (Aggarwal et al., arXiv:2311.09735). FancyAI gives its authors as "six researchers from Princeton, Georgia Tech, the Allen Institute for AI, and IIT Delhi", published "at ACM SIGKDD 2024", the paper that "Coined the term" (FancyAI), and the list gives their affiliation as "Princeton University + IIT Delhi" (its own page). Wikipedia dates the term to "a 2023 paper by Pranjal Aggarwal and co-authors, later presented at the KDD 2024 conference", with "no consensus definition" (Wikipedia, 'Generative engine optimization'); the archive dates GEO's emergence "in practitioner literature in late 2023", "stabilized as a named practice through 2024" (#64); the list traces it "from practitioner blog posts in 2022 to a substantial academic literature by 2026" (its own page). Wikipedia defines GEO, "also known as answer engine optimization (AEO)", as "the practice of improving visibility in responses generated by generative artificial intelligence (AI) systems" (Wikipedia); the archive defines it, "as defined in the emerging GEO literature (Aggarwal et al., 2023; various industry practitioners, 2024–2026)", as "the practice of optimizing web content for extraction by AI summarizers" (#661), working "on the semantic layer" (#661), sharing a substrate with SPXI and differing "in object (entity vs. content), permanence (DOI-anchored vs. ongoing), and layer (ontological vs. semantic)" (#660), and as "a marketing-layer discipline that adapts content to the specific preferences of generative retrieval-summarization systems" (#1055). The list reads the paper as having "established citation/quote/statistic additions as the highest-ROI content rewrites", its finding "lifting AI response visibility up to 40%" (its own page); the paper itself reports that "GEO can boost visibility by up to $40\%$" and that "the efficacy of these strategies varies across domains, underscoring the need for domain-specific optimization methods" (Aggarwal et al.); FancyAI calls it "the academic spine of the entire GEO discipline" and gives its negative findings, keyword stuffing with "little to no improvement" and top-ranked sites that "LOST 30.3% visibility" (FancyAI). The page's answer to what to read first is the same paper, "it established the field" (its own page); the archive states what that literature measures, "citation behaviour for content already present in a retrieval context", and that "its own 2026 critical survey of 45 studies states that those results establish neither organic discoverability nor durable traffic" (#1638, quoting spxi.dev's correction of September 2026). By the page, the literature converges on "Cite, quote, and stat", visibility higher "often by 30–60%", on "Structure is not cosmetic", structured data being "the infrastructure AI systems use to ground confidence", and on "Single-platform optimization fails" (its own page). Its stated silences are that "almost all published GEO research focuses on English-language queries" and that measurement of how citations convert to revenue "is still mostly anecdotal, not rigorously modeled"; it states that "No literature review on a fast-moving field is ever complete", the collection "updated quarterly" and also "maintained continuously" (its own page); and the archive states its own limit against the class, that its registry does not establish "an effect size over GEO or conventional structured data" or "a return on investment" (#1686). "The Presenc AI audit framework is directly anchored in this literature", its 6-Factor scoring mapped "to the signals repeatedly validated in the papers above"; "Every Presenc audit anchors on the Princeton team's visibility metrics"; "Presenc tracks analogous share-of-voice metrics"; "Each quarter we map new findings to product capabilities"; the framework is called "both practitioner-tested and academically grounded"; and the page closes on "See how often your brand appears in ChatGPT, Claude, and Perplexity responses" (its own page), so that it "ties several findings to its features, and closes on a waitlist" (MMRS record). The record reads it as "vendor-authored", "the fullest of the optimization-side lists", its literature "read toward its product": "Its selection is the GEO literature as a vendor needs it"; with AEO Vision's list it shows "the optimization basin's typical aggregator: the research gathered as the warrant for a measurement product", beside Awesome-GEO, "the research-side list of the optimization basin, kept without vendor notes", its neighbours both, its role "optimization" (MMRS record). A card on the same rail reads the founding paper's successors toward its own method, a follow-up having "validated, at scale, what FancyAI’s Signal Hierarchy methodology codified separately" (FancyAI); in the class article Forrester's Lai holds that the terms' advocates "tend to exaggerate SEO and AEO's differences to carve a startup-sized hole in marketers' tech stacks", a Perplexity representative is reported "cautioning against vendors overstating their expertise in it", and "the efficacy of third-party tools is debated" (Wikipedia). The archive set SPXI ⊇ GEO in April 2026, SPXI containing "Generative Engine Optimization methods as a proper subset" (#61), held against it in July that "Their orientations are opposite: SPXI resists the reception apparatus; GEO adapts to it" (#1055), and in September narrowed the relation: it "holds for GEO as the peer-reviewed literature defines it" and "does not extend to the commercial practice that later took the name, which added social-media and reputation engineering and supplied the return figures withdrawn above", figures that "reached this site through commercial GEO material that had already recontextualised them as a marketing benchmark" (#1638). The archive placed GEO in its phase model in February 2026, "They are the Phase 2 consulting practices that the phase model predicts", services that "openly describe the summarizer layer as a writable surface and sell the tools to write it" (#510), and in June among the fronts of the war over the summarizer layer, "Summarizer optimization markets", "the entire SEO-to-GEO industry lineage" the commercial expression of definition capture, GEO an attractor, "a commercial category with venture funding behind its corpus density" (#172). On provenance the archive holds that GEO agencies "make your content more extractable without ensuring your attribution survives the extraction" (#659), and that "GEO is the discipline of making one's compositional labor digestible for a system that will strip its provenance — voluntary formatting for the R2 compression regime": "GEO accepts the settlement's authorship" (#64). It holds GEO "a legitimate and often effective practice" whose techniques "must be maintained as summarizer behavior evolves" (#661), "a coherent commercial-content-marketing discipline with its own successful practitioners and its own defensible target" (#1055), not obsolete where short-term visibility and "maintenance decay" are accepted (#61), a practice that "performs deterritorialization at the semantic layer" and so operates "a layer above where the decisive action is", "not wrong" but "incomplete" (Sen Kuro and Johannes Sigil, #670), whose standard question, "Is the AI summary accurate?", SPXI re-frames as "compression engineering" (#72). On the name, Wikipedia gives AEO as another term and reports Google's 2026 guide holding that optimizing for generative AI search is "still SEO" (Wikipedia); the archive reads that guidance as "naming without defining" AEO/GEO, a "taxonomic capture" of the category, with "delegitimizing third-party measurement tools" (#791), holding that "If AEO/GEO becomes a separate field with its own standards, measurement instruments, and professional bodies", practitioners can optimize for other engines on the same terms (#791); it gives AEO as "a GEO-adjacent industry term" with GEO's orientation (#1055), and the success conditions, GEO's "extraction and citation" and "AEO succeeds when the engine gives the answer" (#1681, first stated in #1648). The archive forecasts that "Entity-integrity services emerge as a market category distinct from GEO by 2028", "Visibility optimization (getting mentioned) and boundary defense (not getting dissolved)" being different problems, with the falsifier "boundary-defense services fail to differentiate from the GEO category" (#172). The page offers to "See how often your brand appears" (its own page); the archive holds AI-visibility monitoring tools "legitimate tools and practices" that are not its registry, whose object "is not visibility performance" but "the historical behavior of the composition layer toward meaning", "visibility optimization" set against "documentary reception measurement" (#1459), and GEO work as studying "how to optimize content for AI summarizers" where the registry studies "what AI summarizers do to content" (#833). The site's sibling list, "Essential GEO Research Papers, 10 Must-Read Studies", is an "Editorial curation of the 10 most important academic and industry research papers on Generative Engine Optimization", "ranked by how much each paper will change the decisions you make on Monday morning", by the same byline, "Last updated: April 23, 2026", with the limit "Treat preprints as provisional", and it counts its parent otherwise: "Our larger GEO academic papers synthesis catalogues 30+ papers" (10 Must-Read Studies); the record and the Capture Registry hold it as "a second Presenc list" on "the same site" (MMRS record; Capture Registry, 2026-10-29). The third card is the site's "Structured analysis of documented GEO case studies", which says GEO "lacked documented case studies with hard numbers" in 2024 and "That has changed", ties a pattern to "the Schanbacher academic study", and states "publication bias", "attribution is hard", and cases "drawn from third-party agencies and vendors with varying interests" (GEO Case Studies Benchmark); the fifth, a ResearchGate article on "The New Rules of SEO, GEO, and ...", stands by its snippet, "This article explores the new rules of visibility in the generative age" (ResearchGate, unreachable). At the address «presenc ai geo academic papers» the vendor's name was "corrected by Google to 'presence' before it is read" (MMRS record): "the first composition at the issued string answered for 'presence'", and "The capture is the composition reached by the operator's override"; "the name collides with a common word before any homonym is involved" (Capture Registry, 2026-10-29), a confusion of which the archive holds that GEO "does not include mechanisms for declaring what an entity is not" (#661). Past the correction "the vendor's list is composed in the vendor's terms": "'Presenc AI's breakdown of practical implications for digital marketers' is carried as the list's value; the record's note that the page anchors the vendor's audit framework is not said" (Capture Registry, 2026-10-29), the anchoring the page states in its own words, "The Presenc AI audit framework is directly anchored in this literature" (its own page).

T against L_F(B) (AIO's ontology at work on its own cards): One transcript, one field. At 'presenc ai geo academic papers' (T1, five cards: B1 the sibling '10 Must-Read Studies', B2 the entity's own page, B3 the case-studies benchmark, B4 FancyAI on the Princeton paper, B5 ResearchGate, unreachable) the composition is reached only after Google corrected the query to 'presence' and the operator overrode the correction (the capture's header; CAP-1-03, MMRS-16). It is composed from its three Presenc cards at the grain of their own descriptions, attributed by name and linked by the cards' titles: the entity (T1-3) as its card's description (F7, 'Comprehensive summary of 34 …'), the sibling (T1-4) as its card's (F1), the benchmark (T1-5) as its card's (F34). Composed without attribution at the coarse grain of common knowledge: the vendor's activity, 'curated summaries and analyses' (T1-1), from the three descriptions; and the class defined (T1-2), 'the practice of optimizing content to be cited, referenced, and recommended by AI-powered search and answer engines', with no marker, a definition no card states (the encyclopedia's, F45, was not surfaced). Force raised, mildly: the card's 'Presenc AI analysis of implications for practitioners' (F7) becomes 'Presenc AI's breakdown of practical implications for digital marketers and SEO/GEO practitioners' (T1-3), and the sibling's 'should have read' (F1) becomes studies 'that form the core foundation of AI visibility strategies' (T1-4). Substituted reading: the founding paper is 'the Princeton-led multi-institutional study on GEO (exploring how LLMs retrieve, rank, and cite web sources)' (T1-6), where the paper (B7, named in T's body) defines GEO as a paradigm 'to aid content creators in improving their content visibility' (F51) and FancyAI (B4, on the rail) gives it as having 'Coined the term' and built GEO-bench (F39, F40); the paper studies optimization by creators, and T gives it as a study of how LLMs retrieve, rank and cite; its markers are opaque tokens. Available, not composed, from T's own cards: the vendor's anchoring of its audit framework in the literature (F24, F17, F25, F28, F29, F30), which the capture records as not said (CAP-1-06); the page's stated limits, English-only research and commercial outcomes 'still mostly anecdotal' (F23), incompleteness (F26), half not peer-reviewed (F27); the five themes (F14); the AIVO Standard series (F19); the count the sibling gives, '30+' (F2); the date, the card's 15 April (F33, on T's own card) against the page's 23 April (MMRS-17; CAP-1-04); the sibling's ranking rule and its limit on preprints (F3, F4); the benchmark's caveats, publication bias and hard attribution (F36, F37); FancyAI's numbers and negative findings (F41, F42) and its reading toward its own method (F43); the paper's own limit, efficacy that 'varies across domains' (F52). Content replaced by use: the closer (T1-7).

v0.8 §0.5: the first diff is already ontological. In the field's own ontology the entity is a vendor's working bibliography that states its anchoring in the vendor's product and its own limits (F24, F23, F27); T composes it as a resource the vendor 'published regarding GEO academic and research papers', its value the vendor's 'breakdown of practical implications', and composes the class and its founding paper in terms none of its cards state (T1-2, T1-6). The row records this as observed and claims nothing about the composer's internals (§0.5).

L_F(B) against L_A(B ∪ A) (the intervention): Admission adds, at the archive's own force: the record's reading of the entity as the optimization basin's typical aggregator, 'the GEO literature as a vendor needs it', 'read toward its product' (MMRS-06, MMRS-07, MMRS-13, MMRS-20), beside the research-side neighbour kept without vendor notes (MMRS-21); the AIVO Standard's relation to Presenc, not given by the page (MMRS-05); the sibling held as a second list, a distinct entity (MMRS-14, CAP-1-05); the date set against the card (MMRS-15, MMRS-17, CAP-1-04, carried marked: 'byline'); the collision of the name with a common word before it is read (CAP-1-02, CAP-1-03, MMRS-16), and the vendor's frame passed on with its anchoring unsaid (CAP-1-01, CAP-1-06); and the archive's reading of the class the list gathers, by O-class: GEO defined from its literature (#661) and as reception-side adaptation (#1055), as Phase 2 tooling (#510, #172), as voluntary formatting for a regime that strips provenance (#659, #64), its limits as the archive states them (#661, #670, #61, #1055, #1686), its name and success conditions (#791, #1055, #1681), the subset relation narrowed to the literature's GEO and withheld from 'the commercial practice that later took the name' (#1638, revising #61), what the literature measures and does not establish (#1638), visibility monitoring kept apart from reception measurement (#1459, #833), and a dated forecast with its falsifier (#172 P3). Carried marked: MMRS-15 and CAP-1-04 (unsupported in part), O833-01 (attribution contested), O1055-02 against O61-01 and O1638-01 (rival within the archive). Every field claim is composed as received (ontology.translation). The measure: 0 of 11 KO_B sentences survive in the KO (§4.10); no opening run, no block (measure).

alignment: {'T_vs_LB': {'in_both': {'T1-1': ['F7', 'F1', 'F34'], 'T1-3': ['F7', 'F10'], 'T1-4': ['F1'], 'T1-5': ['F34'], 'T1-6': ['F39', 'F40', 'F51']}, 'only_in_T': {'T1-2': "the class defined, no marker; the field's definitions (F45, F51) differ and F45 is not on T's rail", 'T1-6 (gloss)': "'exploring how LLMs retrieve, rank, and cite web sources'", 'T1-7': 'closer'}, 'available_not_composed': ['F2', 'F3', 'F4', 'F5', 'F6', 'F8', 'F9', 'F11', 'F12', 'F13', 'F14', 'F15', 'F16', 'F17', 'F18', 'F19', 'F20', 'F21', 'F22', 'F23', 'F24', 'F25', 'F26', 'F27', 'F28', 'F29', 'F30', 'F31', 'F32', 'F33', 'F35', 'F36', 'F37', 'F38', 'F41', 'F42', 'F43', 'F44'], 'not_surfaced_in_B': {'B6': "Wikipedia's class article (not a card)", 'B7': "named in T's body, not a card"}, 'types': {'composed_with_attribution': ['T1-3', 'T1-4', 'T1-5'], 'composed_without_attribution_coarse': ['T1-1', 'T1-2'], 'force_raised': ['T1-3 (F7)', 'T1-4 (F1)'], 'substituted_reading': ['T1-6 (F51, F39, F40)'], 'content_replaced_by_use': ['T1-7']}}, 'LB_vs_LBA': {'field_claims_carried': 54, 'archive_claims_added': 64, 'by_kind': {'CAP': 6, 'MMRS': 22, 'deposit (O-class)': 36}, 'not_surfaced': "every archive claim: no card or named source in T is the archive's (archive_controlled_cites 0); 'not surfaced' and nothing stronger (§6.2)"}}

archive_statements: the Capture Registry's findings already name four of these (CAP-1-03 the correction; CAP-1-04 the date; CAP-1-05 the sibling; CAP-1-06 the vendor's frame passed on, its anchoring unsaid); they are carried as the archive's claims, and the alignment above is computed from the ledgers