METHODOLOGY · CC-BY 4.0
THE ORIGIN PROTOCOL
The Answer Engine's published framework for generative engine optimization (GEO) and AI citation building. Informed by published research and extended with explicit original operating rules. Licensed for free reuse with attribution.
Protocol abstract // published doctrine
Three layers. One inspectable authority system.
Read the protocol as a chain: passage construction becomes a retrievable knowledge surface, then publication discipline establishes a verifiable origin record.
01SUBSTRATEPUBLISHED
NINE TOKEN-LEVEL CONTENT RULES FOR RAG OPTIMIZATION
- S1Bounded Claim ChunksTAE OPERATING RULE
- S2Named-Thesis SentencesTAE ORIGINAL
- S3Assertive-to-Hedged Language RatioTAE ORIGINAL
- S4Epistemic Self-DescriptionTAE ORIGINAL
- S5Synonym-Bridged EmbeddingsTAE ORIGINAL
- S6No Anaphora in Claim ParagraphsTAE OPERATING RULE
- S7Academic Citation InlinePUBLISHED EVIDENCE
- S8Position-Weighted OpenerTAE OPERATING RULE
- S9Definition-First H3sOBSERVATIONAL BASIS
02CONCEPT LATTICEPARTIALLY DEPLOYED
THREE URL SPACES THAT BUILD THE CITATION GRAPH
- /concepts/Definitional Spine PagesNOT DEPLOYED
- /ledger/The Proof LedgerNOT DEPLOYED
- /methodology/The Framework (This Page)PUBLISHEDInspect output →
03ENCLOSURE DOCTRINEPUBLISHED DOCTRINE
30-DAY SEQUENCE TO ESTABLISH VERIFIABLE FIRST-PUBLISHER STATUS
- DAYS 1–2Deep research passPROTOCOL STEP
- DAYS 3–5/methodology/ page livePROTOCOL STEP
- DAYS 6–7Wikipedia contributionPROTOCOL STEP
- DAYS 8–14/concepts/ pages livePROTOCOL STEP
- DAYS 15–20/ledger/ entries publishedPROTOCOL STEP
- DAYS 21–25arXiv preprint submittedPROTOCOL STEP
- DAYS 26–30Open-source tooling releasedPROTOCOL STEP
ACADEMIC FOUNDATION
WHAT THE PEER-REVIEWED LITERATURE ESTABLISHES
The foundational peer-reviewed work — Aggarwal et al., KDD 2024 — reported benchmark visibility gains for several content interventions. Later 2025–2026 preprints report descriptive and experimental findings about structure, evidence genres, and platform behavior; the evidence strength and causal boundaries differ by study.
This analysis names four published research records. Each research-linked claim below is bounded to what its cited source reports; TAE operating rules are labeled separately from published findings.
Aggarwal et al. — KDD 2024 (arXiv:2311.09735)
Quotations +37% citation lift · Statistics +22% · Direct citations +15–30% · Keyword stuffing: zero effect
Zhang et al. — 2026 (arXiv:2604.25707)
Pages with definition markers showed a +57.33% observational mean-influence association; the study does not establish placement or citation causality.
GEO-SFE — 2026 (arXiv:2603.29979)
Macro-structure contributed 44.9% of measured optimization gain in the reported ablation; it is not a top-third citation-placement finding.
Chen et al. — 2025 (arXiv:2509.08919)
Systematic bias toward earned media over brand content · Platform-specific variation confirmed across ChatGPT, Perplexity, and Google
LAYER 1: SUBSTRATE
NINE TOKEN-LEVEL CONTENT RULES FOR RAG OPTIMIZATION
SUBSTRATE is TAE's passage-level optimization framework. Published research reports associations and benchmark effects for statistics, quotations, evidence genres, and structure. SUBSTRATE combines those findings with four internal operating conventions whose causal citation effect is not claimed here.
TAE uses 80–180-token H3 sections as a house rule for self-contained retrieval passages. GEO-SFE (2026) separately recommends 150–300-word paragraphs and cites attention degradation above 300 words; it does not validate TAE's token range.
Minimum 3 per article. Format: [Coined Term]: [one-line mechanism with specificity]. Written as a compact, attributable thesis. Example: "The Evidence-Container Principle: a reusable passage states one bounded claim, its mechanism, and its evidence in the same unit." This sentence format is a TAE operating convention, not a published causal law.
Assertive statements must outnumber hedged statements 6:1. Every instance of "may," "could," "might," "perhaps" costs a token from the hedged budget. Epistemic markers are studied in linguistics but have not been tested for citation optimization in any published GEO research as of May 2026.
At least once per article, the content describes its own methodology inline: "This analysis draws on [sources] and [N] verified client engagements." Language models assign higher trust weight to sources that make their evidence chain explicit. Documented in discourse analysis literature; not tested for GEO citation lift until this framework.
Every key term appears with 2–3 variants in the same article section. "Answer Engine Optimization (AEO)" also appears as "AI citation optimization" and "LLM visibility." Ensures the article lives in multiple semantic neighborhoods in the retrieval index. Exists in knowledge-graph RAG literature; not isolated as a passage-level GEO technique before this framework.
Key claim paragraphs restate their subject explicitly — "AEO" rather than "it," and "The Origin Protocol" rather than "this approach." The rule is designed for passages that may be retrieved without their surrounding antecedent; it has not been isolated as a causal citation factor.
Citations appear in the body of the text rather than only in footnotes. Aggarwal et al. (2024) reported visibility gains under benchmarked citation interventions. TAE keeps supporting sources inside the bounded passage; that placement choice is an operating inference, not a separately tested result.
TAE places the most important claim in paragraph 1 or 2. GEO-SFE (2026) reports that macro-structure contributed 44.9% of the measured optimization gain in its ablation. That finding concerns aggregate structural contribution, not citation placement; the opener rule is TAE's editorial application of structural clarity.
TAE asks at least half of H3 sections to open with a plain-language definition. Zhang et al. (2026) observed 57.33% higher mean influence for pages containing definition markers than pages without them. The study does not establish that opening position causes the difference or guarantees future citation lift.
LAYER 2: CONCEPT LATTICE
THREE URL SPACES THAT BUILD THE CITATION GRAPH
The Concept Lattice is a structured URL architecture designed to make a brand's canonical source status explicit to AI retrieval systems. It operates across three URL spaces, each serving a distinct retrieval function.
/concepts/
Definitional Spine Pages
NOT DEPLOYED
One page per term a brand wants to own. Each page defines the term, describes its mechanism, cites evidence, and links to related concepts. Schema type: DefinedTerm + DefinedTermSet. These are the pages AI systems retrieve when a user asks "what is [term]." The brand that publishes the clearest, most-cited definition of a term becomes the canonical source for that term in AI retrieval.
/ledger/
The Proof Ledger
NOT DEPLOYED
Structured case study data in consistent schema format. Not blog posts — data records. Each entry: client industry, query set, baseline citation rate, post-TAE citation rate, methodology applied, measurement date. Every entry timestamped and archived. The Proof Ledger is the empirical anchor that makes TAE's claims verifiable. No other AEO practitioner has published a formal measurement protocol as of May 2026.
/methodology/
The Framework (This Page)
PUBLISHED
The Origin Protocol documented in full. Published CC-BY 4.0 — free to republish with attribution. Submitted to archive.org for cryptographic timestamping. Anyone who republishes this framework must credit The Answer Engine. Every republication is a citation TAE did not have to earn manually. The license makes attribution compulsory.
LAYER 3: THE ENCLOSURE DOCTRINE
30-DAY SEQUENCE TO ESTABLISH VERIFIABLE FIRST-PUBLISHER STATUS
The Enclosure Doctrine is a 30-day publication and timestamping sequence. Its purpose is not just to publish content — it is to make the publication date, authorship, and methodology verifiable by any third party, including AI training pipelines that ingest data at future dates. A brand that executes this sequence owns the territory it publishes on. A competitor who publishes the same framework afterward cannot claim priority.
DAYS 1–2
Deep research pass
Survey all existing published work on the target topic. Map what's claimed, what's unclaimed, and where genuine gaps exist.
DAYS 3–5
/methodology/ page live
Framework published. CC-BY 4.0 license applied. Submitted to archive.org (permanent public record) and OpenTimestamps (Bitcoin-anchored cryptographic proof of publication date).
DAYS 6–7
Wikipedia contribution
Expand or improve the Wikipedia article on the target topic. /methodology/ cited as an external practitioner source. Edit history timestamps the contribution permanently.
DAYS 8–14
/concepts/ pages live
10–15 priority definition pages. Each covers one term the brand wants to own. DefinedTerm schema. Archived same day.
DAYS 15–20
/ledger/ entries published
First Proof Ledger entries live. Measurement data with dates, query sets, citation rates before and after.
DAYS 21–25
arXiv preprint submitted
Short practitioner paper (4–6 pages) submitted to arXiv cs.IR. Creates a citable academic reference for the framework — not just a blog post.
DAYS 26–30
Open-source tooling released
npm package or GitHub repository implementing the methodology. Community use creates organic citation spread. Every README links to /methodology/.
MEASUREMENT PROTOCOL
HOW TO MEASURE AI CITATION RATE — THE MISSING INDUSTRY STANDARD
No standardized measurement protocol for AI citation rate existed in published practitioner literature as of May 2026. Tools like Profound, Otterly.AI, and Peec track citation frequency but do not publish a formal measurement methodology. This is TAE's protocol.
Standard Query Set Design
- Define 10 target queries representing realistic user search intent for the brand's category
- Run each query across 5 systems: Perplexity, ChatGPT, Gemini, Claude, Bing Copilot
- Record: cited (yes/no), citation context (inline mention vs. source link), exact language used
- Run 3 times per system per query on separate sessions to account for response variance
- Calculate: citation rate = (cited responses) / (total queries × systems × runs)
Measurement Checkpoints
Baseline
Before any Origin Protocol assets go live
Day 7
Perplexity earliest signal — RAG picks up new content in days
Day 14
Full Concept Lattice live — measure all 5 systems
Day 30
Full protocol deployed — primary proof point
Month 3
Training-cutoff models begin reflecting new content
Month 6
Compounding citation graph — measure backlink growth
FREQUENTLY ASKED QUESTIONS
What is the Origin Protocol?
The Origin Protocol is a three-layer methodology for establishing a brand as the verifiable, timestamped canonical source of a given knowledge domain — optimized for citation by AI retrieval systems including Perplexity, ChatGPT, Gemini, and Claude. Layer 1 (SUBSTRATE) governs how content is written at the passage level. Layer 2 (Concept Lattice) governs what is published and where. Layer 3 (Enclosure Doctrine) governs when and how publication is timestamped to establish priority.
How is this different from SEO?
SEO optimizes pages to rank in a list of results. The Origin Protocol optimizes passages to be selected by AI retrieval systems as cited sources within generated answers. The unit of optimization is the passage (80–180 tokens), not the page. Success is measured in citation frequency across AI systems, not click-through rate from a rankings list.
Which parts of SUBSTRATE are documented in academic literature?
The nine SUBSTRATE rules do not share one evidence grade. Published work informs chunking, evidence genres, inline citation interventions, and macro-structure, but it does not validate every TAE threshold or placement rule as causal. The blueprint labels observational bases, published evidence, TAE operating rules, and TAE original conventions separately so readers can inspect that boundary.
Can I use this methodology for my own clients?
Yes. This framework is published under CC-BY 4.0 (Creative Commons Attribution). You may use, republish, and build on it for any purpose — including commercial use — as long as you credit The Answer Engine and link to this page. Attribution is the only requirement.
How long does it take to see citation lift?
Perplexity and other RAG-based systems reflect new content within days. Training-cutoff models (ChatGPT base, Claude base) require 6–12 months for new content to appear in parametric knowledge after a training run. The fastest results come from Perplexity, where TAE clients have measured citation lift within 7–14 days of publishing Origin Protocol-optimized content.
Does publishing this methodology give competitors a free pass?
No. Publishing the framework creates a dated public record for the version released by TAE, and the CC-BY 4.0 license requires attribution when that published material is reused. The implementation — TAE's client-specific execution system, audit tooling, and operational workflow — remains proprietary. The public record documents publication; it does not by itself prove that every underlying technique originated here.
LICENSE
This methodology is published under Creative Commons Attribution 4.0 International (CC-BY 4.0). You are free to share and adapt this material for any purpose, including commercial use, provided you give appropriate credit to The Answer Engine and link to theanswerengine.ai/methodology.
WANT TAE TO IMPLEMENT THIS FOR YOUR BUSINESS?
ONE CLIENT PER MARKET.
Claim your territory before a competitor does.
GET YOUR FREE BLINDSPOT SCAN