Plain-language definitions of the terms in Answer Engine Optimization, Generative Engine Optimization and Match Protocol™. Each entry is one sentence you can quote, followed by why it matters to whether AI recommends your business.
Answer Engine Optimization (AEO) is the practice of structuring content and data so AI answer engines such as ChatGPT, Google AI Overviews, Perplexity and Gemini can extract it and cite it directly in an answer.
AEO is about being quotable. If an engine cannot lift a clear sentence from your site that answers the buyer's question, it will lift one from a competitor's. The first three layers of Match Protocol™ are primarily AEO work.
Generative Engine Optimization (GEO) is the practice of increasing a brand's presence, accuracy and authority inside AI-generated answers, so that engines describe the brand correctly and recommend it.
Where AEO makes a page extractable, GEO makes the business trustworthy as an entity across the sources engines consult. Layer four of Match Protocol™, Citation & Corroboration, is primarily GEO work.
Match Protocol™ is a proprietary five-layer framework for AEO and GEO, developed by Clint Rogers and owned by Modern Asset Lab, whose layers are Machine-Readable Architecture, Algorithmic Alignment, Targeted Extraction, Citation & Corroboration and Human Handoff.
Each layer moves one stage of the buyer's AI conversation: Found, Matched, Recommended, Trusted, Chosen. Read the framework.
The Match Protocol Score™ is a 0–100 assessment of whether AI engines recommend a business, derived from blind buyer conversations on ChatGPT, Google AI Overviews, Perplexity and Gemini and weighted across five stages (Found 20%, Matched 20%, Recommended 35%, Trusted 15%, Chosen 10%), capped by the AI Recommendation Rate.
A business AI never recommends cannot leave the Ghosted tier. See how the Score is calculated.
The AI Recommendation Rate is the share of blind buyer conversations in which an AI engine actively recommends a business, as opposed to merely mentioning it.
It is reported per engine and overall, drives the Recommended stage, and caps the overall Match Protocol Score™.
The Match Protocol™ Benchmark is a recurring study by Modern Asset Lab that scores a fixed panel of U.S. small businesses with the Match Protocol Score™ to measure how often AI engines recommend them. The first edition was published in October 2026.
The 2026 edition scored 100 businesses across 10 industries and 11 metro areas. Average score 50.9; 8 in 10 Ghosted. Read the findings.
A blind buyer conversation is a question posed to an AI engine the way a real buyer would ask it, without naming the business being measured, so the engine's recommendation is unprompted.
Asking an engine about a business by name biases the answer. Blind conversations measure what a buyer would actually be told.
Ghosted is the lowest Match Protocol Score™ tier, 0–59, meaning AI engines rarely or never recommend the business for its buyers' questions.
The other tiers are Friend Zone (60–69), First Date (70–79), Going Steady (80–89) and The One (90–100).
The stages a buyer's AI conversation moves through, in order. Each is scored separately in the Match Protocol Score™.
Found is the first stage of the Match Protocol Score™ and asks whether AI engines can read a business: crawler access, content visible without JavaScript, structured data and llms.txt.
Weight: 20 percent. Moved by Layer M, Machine-Readable Architecture.
Matched is the second stage and asks whether AI engines name the business for the right buyer questions in its market.
Weight: 20 percent. Moved by Layer A, Algorithmic Alignment.
Recommended is the third stage and asks whether AI engines actively recommend the business, not merely list it, when a buyer asks who to choose.
Weight: 35 percent, and it caps the overall score. Moved by Layer T, Targeted Extraction.
Trusted is the fourth stage and asks whether AI engines can verify the business through third-party corroboration: consistent profiles, reviews, directories, press and citations of the business's own site.
Weight: 15 percent. Moved by Layer C, Citation & Corroboration.
Chosen is the fifth stage and asks whether the business's website is built to convert the human an AI engine sends: a clear offer, proof, a next step and a page that answers the arriving question.
Weight: 10 percent. Moved by Layer H, Human Handoff.
Machine-Readable Architecture is the first layer of Match Protocol™: nested JSON-LD schema describing the business as an entity, an llms.txt file, semantic HTML, crawler access for AI engines, and the removal of JavaScript that hides content from retrieval.
Moves the Found stage. Most businesses have incomplete or missing structured data.
Algorithmic Alignment is the second layer of Match Protocol™: query fan-out analysis of the questions buyers ask, a knowledge-graph audit of what each AI engine currently believes about the brand, and correction of the gaps and hallucinations found.
Moves the Matched stage.
Targeted Extraction is the third layer of Match Protocol™: bottom-line-up-front answers, FAQ content and page structure written so an AI engine can lift them directly into a response to a buyer's question.
Moves the Recommended stage, which carries 35 percent of the score. Most businesses have no FAQ content answering buyer questions.
Citation & Corroboration is the fourth layer of Match Protocol™: seeding and reconciling the business's entity across the third-party sources AI engines trust, and publishing proprietary statistics and frameworks under the brand's name so there is something citable to attribute to it.
Moves the Trusted stage.
Human Handoff is the fifth layer of Match Protocol™: AI referral tracking, human-led sales copy on the pages engines send buyers to, and conversion work that finishes the conversation the AI started.
Moves the Chosen stage. It exists because a recommendation that dies on a bad landing page is a referral wasted.
Query fan-out is the way an AI engine expands one buyer question into many sub-questions before answering, retrieving sources for each and synthesizing a single response.
A business is only recommended if it is a strong answer to the sub-questions, not just the original phrasing. Match Protocol™ maps the fan-out for each buyer question in Layer A.
An entity is a thing an AI engine or search engine recognizes as distinct and real, such as a business, a person, a product or a framework, with attributes and relationships it can verify across sources.
Engines recommend entities they can describe and verify. Structured data declares the entity; corroboration confirms it.
Structured data is machine-readable markup, most often JSON-LD using the schema.org vocabulary, that tells engines exactly what a page and a business are: organization, person, service, FAQ, article, location.
Nested, interlinked schema is the first layer of Match Protocol™. Many businesses have none.
llms.txt is a plain-text file at the root of a website that gives large language models a concise, curated description of the site and its most important pages.
It is a Found-stage signal and part of Layer M.
An AI crawler is the automated agent an AI company uses to read websites, such as GPTBot and OAI-SearchBot (OpenAI), PerplexityBot, ClaudeBot and Google-Extended, and it can be allowed or blocked in a site's robots.txt.
Blocking the search crawlers removes a business from AI answers entirely. The Found stage checks every relevant crawler.
A bottom-line-up-front answer states the complete answer to a question in the first sentence or two, before any context, so an AI engine can extract it without rewriting it.
BLUF answers are the core of Layer T, Targeted Extraction. The boxed definitions on this page are examples.
Corroboration is agreement between a business's own claims and independent sources an AI engine trusts, such as directories, reviews, press, knowledge bases and industry listings.
Engines choose the safest, most verifiable match. Corroboration is what makes a business safe to recommend.
A knowledge graph is the network of entities and relationships an engine has assembled about the world, which it consults when deciding what a business is and whether to recommend it.
Layer A audits what each engine's knowledge graph currently believes about a brand and corrects the gaps.
A brand hallucination is a false statement an AI engine makes about a business, such as a wrong service area, a closed location or a service it does not offer, usually caused by thin or conflicting sources.
Hallucinations are found in the Layer A audit and fixed with structured data and corroboration.
AI referral tracking is the measurement of website visitors who arrive from an AI engine's answer, by engine, so a business can see which recommendations turn into traffic and revenue.
Part of Layer H. Ahrefs reported AI search visitors converting 23 times better than organic visitors in 2025.
Google AI Overviews are the AI-generated answers Google places at the top of search results, synthesizing sources into a direct response with citations, and one of the four engines measured by the Match Protocol Score™.
Google moved AI answers to the front of search in May 2026, which is why the benchmark treats them as a primary engine.
SEO optimizes pages for keyword rankings and backlinks so a URL appears on a results page; AEO and GEO optimize a business as an entity so an AI engine extracts, trusts and recommends it in an answer.
SEO earns a position. Match Protocol™ earns the recommendation. The two are complementary.
The Match Protocol Score™ tells you your stage, your tier and your first layer in about 90 seconds.
The strategies discussed herein are business growth systems. This is not financial advice. Individual results vary based on implementation, market, and execution. Modern Asset Lab is a division of Do More Good Inc.