AVE Studio · AI Visibility Engineering
The language we use to make businesses legible to AI — defined precisely, in one place. GEO and AEO are the field's terms; KEO, Model Mindshare, M2M and B2M are frameworks we propose and define here.
Being found, included and cited when an AI generates an answer — at answer time, from the live web.
Being found, included and cited when an AI generates an answer — at answer time, from the live web. GEO is the retrieval-side work of AI visibility: access, structure and machine-readable claims, so answer engines select and name the brand. It is measured by repeated sampling across surfaces, with confidence ranges — never from a single query.
Being the extracted answer to a specific question — at answer time.
Being the extracted answer to a specific question — at answer time. AEO is the featured-answer genre (featured snippets, AI Overviews): one well-structured, self-contained block gets lifted into the answer box. It differs from classic SEO, which ranks a list of links — and from GEO, which works toward being named inside a generated answer.
The verified, canonical, readable, citable record AI models learn from — so each new model is trained to know and recall your business correctly.
The verified, canonical, readable, citable record AI models learn from — so each new model is trained to know and recall your business correctly. KEO is the generation after GEO: GEO optimizes what engines retrieve at answer time, KEO builds what models remember offline: entity records, corpus presence and consistent claims — verified by closed-book measurement at each model release.
framework proposed by AVE StudioWhat is Knowledge Engine Optimization? →
The share of an AI model's recall a brand holds in its category — a measured signal with a confidence range, not a guarantee. Measured closed-book (with web search off) as a training-time, parametric signal — the generation after retrieval-time GEO visibility.
framework proposed by AVE StudioWhat is Model Mindshare? →
Marketing aimed at the machines that now mediate discovery, not only at people.
Marketing aimed at the machines that now mediate discovery, not only at people. M2M treats AI assistants, answer engines and their crawlers as an audience with reading rules of their own: structured claims, verifiable entities, consistent facts — because a growing share of buying decisions is shaped by what machines find, remember and say.
framework proposed by AVE Studio
The transaction tier of AI visibility: being bought through an AI agent that acts on a person's behalf, via emerging agentic-commerce standards. A framing AVE Studio uses for the third tier (be seen → be known → be bought); the underlying agentic-commerce protocols are industry standards, not AVE's.
framework proposed by AVE StudioWhat is Business-to-machine? →
Machine equity is the business asset that arises from AI systems being able to recognize, understand, recall, recommend — and, on a user's behalf, transact with — a brand. Composed of three layers: rented reach (retrieval visibility), owned memory (trained model recall), and machine buyability (agent-side transactability). Measured honestly only as a portfolio of ranged, per-layer metrics — never as a single blended score.
framework proposed by AVE StudioWhat is Machine equity? →
The retrieval-side component of machine equity: whether AI systems find and name a brand in live answers. Re-decided at every query — valuable but never owned, like a well-placed rented billboard. Built and measured on the GEO/AEO tier.
framework proposed by AVE Studio
The parametric component of machine equity: what a model knows about a brand from trained memory alone, with no internet access — measured as unaided recall per model and language, always with confidence ranges. Built and measured on the KEO tier; changes only on training-generation cycles (4–16 month lag).
framework proposed by AVE Studio
The transaction-side component of machine equity: whether an AI agent can parse, compare, and complete a purchase from a brand's catalog and checkout — structured product data, agent-payment protocol readiness, machine-readable offers. The AI-transactions (B2M) tier.
framework proposed by AVE Studio
The baseline measurement of AI visibility: week by week it measures how often each AI assistant names you in answers to your category's real buyer questions — per engine, with repeated sampling and confidence ranges. This is the baseline every other module builds on: the effect of every later fix becomes measurable against it.
AVE Studio KEO moduleWhat is Citation Tracker? →
The yardstick-building module: it assembles and measurement-validates the set of real buyer questions — transactional, comparative, educational, informational and branded — that your market actually asks AI systems. Only questions where naming is genuinely decided make the set; tracking and gap diagnosis build on this portfolio.
AVE Studio KEO moduleWhat is Selector? →
The gap-diagnosis module: page by page and question by question it shows why the AI passes you over — which of your pages it compared, and which technical barrier or content gap sits behind every question where you are invisible. Not generic advice: every finding is backed by the actual AI answer as evidence.
AVE Studio KEO moduleWhat is Gap Analysis? →
The fix-delivery module: it turns the diagnosed gaps into a ranked impact-by-effort worklist and ships ready-to-paste fixes — structured data (schema code), FAQ blocks, an AI-crawler steering file and a developer task sheet. Finished work instead of advice: your team can execute it without further interpretation.
AVE Studio KEO moduleWhat is Gap Closure? →
A closed-book measurement of what AI models recall about a brand from memory, with web search off — unaided then aided, per engine and per language. Each dimension carries a confidence interval; no composite score is reported until weights are calibrated. The training-time, parametric counterpart to retrieval-time visibility.
AVE Studio KEO moduleWhat is Model Memory Audit? →
The work of making a brand one clean, verifiable entity for machines — a Wikidata item, a consistent sameAs web and a single canonical fact set across high-authority surfaces — plus source-level correction of false parametric "memories" so the next model generation learns the fix.
AVE Studio KEO moduleWhat is Entity Anchor? →
The input-side program that builds a brand's presence where models actually learn — licensed press, earned community surfaces and a content calendar timed to training cutoffs. It shapes what future corpora contain; it never claims guaranteed inclusion in any model's training data.
AVE Studio KEO moduleWhat is Corpus Campaign? →
Re-running the closed-book battery at each major model release to measure the recall delta the new generation carries — engine by engine, language by language, with a competitor panel as the counterfactual. The honest, after-the-fact evidence that knowledge-engine work moved the model's memory.
AVE Studio KEO moduleWhat is Release Audit? →
Without the internet — from the model's trained memory only, web search off.
Without the internet — from the model's trained memory only, web search off. Closed-book measurement shows what an AI knows about you when it does not search: that knowledge is fixed at training time, persists across a model generation, and cannot be fixed with quick web content — only through presence in the sources models learn from.
Unaided recall: does the model name you without being prompted with your name.
Unaided recall: does the model name you without being prompted with your name. It is the strongest signal of machine memory — the machine counterpart of spontaneous brand awareness: if the model says your name on its own, it genuinely links you to your category. Measured with repeated sampling, per model and language, with confidence ranges.
Aided recall: does the model confirm you when your name is offered.
Aided recall: does the model confirm you when your name is offered. The machine counterpart of prompted brand awareness — a weaker signal than unaided recall, but it shows whether the model knows you at all: can it make accurate statements about you once given the name. The gap between the two is the memory gap itself.
Confidence range: how certain the number is — the statistical uncertainty from the sample.
Confidence range: how certain the number is — the statistical uncertainty that comes from the sample. AI answers vary from run to run, so a single query is a snapshot, not a measurement; every number we report comes from repeated sampling, and the range shows how much is real position versus random fluctuation.
The model's trained memory — what it holds in its weights, not what it retrieves live.
The model's trained memory — what it holds in its weights, not what it retrieves live. Parametric knowledge is fixed at training time from the learning corpus and stays unchanged for the life of the model generation: it cannot be edited with web content afterwards, only through corpus presence before the next training run.
Corpus footprint: how much trace of you exists where models learn — Common Crawl, Wikipedia/Wikidata, mention volume.
Corpus footprint: how much trace of you exists where models learn — Common Crawl, Wikipedia and Wikidata, licensed outlets, mention volume. What counts is not the size of your website but your machine-reachable presence in independent sources: this is what the next model generation's memory of you is built from.
Ghost citation: the AI cites you as a source but never says your name — varies by model.
Ghost citation: the AI cites you as a source but never says your name — your content does the work while your brand stays invisible. It varies by model, and that is what makes it telling: the machine finds and uses you but does not connect you to your entity — typically a symptom of weak entity anchoring.
framework proposed by AVE Studio
Visitors arriving from an AI conversation in which the purchase decision has largely already been made — the behavioral product of an invested research session (sunk cost, commitment, co-production).
Pre-converted traffic: visitors arriving from an AI conversation in which the purchase decision has largely already been made. The sunk cost and co-production of a long research dialogue create commitment: someone clicking through like this comes not to browse but to confirm and complete — which is why it converts exceptionally well.
framework proposed by AVE Studio
How often AI names you when your buyers ask.
Naming rate: how often AI names you when your buyers ask. The headline metric of GEO measurement — computed per question and per engine from repeated sampling, with a confidence range. It does not measure whether you rank: it measures whether your company is named inside the generated answer, where the decision is made.
How much of your site an AI crawler can actually read.
Machine readability: how much of your site an AI crawler can actually read. It is not about design but access and structure: does the page load without JavaScript, is there structured data, are the claims stated unambiguously. What a machine cannot read is also missing from AI answers.
One of the market's names for work aimed at getting a brand to appear in large language model answers.
Search optimization: work whose unit of success is position in a results list and the clicks that follow from it.
An ambiguous acronym: it denotes Google's generated summary, and also general AI optimization.
An umbrella term covering classic SEO, GEO, AEO and KEO — being visible both in result lists and in AI answers.