The AEO/GEO Glossary for Marketing Managers

AEO (Answer Engine Optimisation): What Is It and How Does It Work?

The basics

The practice of optimising your content to be cited in AI-generated answers, rather than just ranked in traditional search results.

Traditional SEO gets you on the list of results. AEO gets you into the answer itself.

The mechanics

When someone asks an AI system a question (whether that’s ChatGPT, Perplexity, or Google’s own AI Overviews) the system doesn’t just search for keywords. It retrieves content from across the web, evaluates it for factual density, structural clarity, and cross-web consensus, then synthesises a single answer from the sources it trusts most.

The sources that get cited tend to share common traits. They provide clear, complete answers to specific questions. They’re structured so AI can easily extract key information. They’re validated by mentions and references across multiple trusted platforms.

Traditional ranking signals like domain authority and backlinks still matter for SEO, but AEO adds a new layer: entity association, third-party validation, and content that’s built to be extracted into an AI-generated response.

Why it matters

You now need two parallel strategies: traditional SEO for rankings, and AEO for citation in AI-generated answers.

This isn’t about choosing Google versus AI. Google itself is becoming an answer engine through AI Overviews. When AI Overviews appear, click-through rates drop from 15% to 8%. Only 12% of AI Overviews link to the #1 ranking result.

Being ranked isn’t enough if you’re not being cited.

Start by auditing whether AI crawlers can actually access your site, reviewing your content structure for question-based headers and modular answers, and checking whether your brand is being mentioned across trusted third-party platforms.

Related terms: GEO, Entity Association, Structural Clarity, AI Overviews, Zero-Click Searches

AI Crawlers: What Are They and Why Do They Matter?

In plain English

Automated bots operated by AI companies that visit your website to read and index your content.

They’re the AI equivalent of Googlebot. If they can’t access your site, AI systems can’t cite your content in their answers.

Under the hood

Each major AI platform runs its own crawler. GPTBot and OAI-SearchBot are OpenAI’s (for ChatGPT), ClaudeBot is Anthropic’s (for Claude), and PerplexityBot is Perplexity’s. Google’s existing Googlebot handles content for AI Overviews as well as traditional search.

These bots identify themselves through user-agent strings in your server logs. They respect robots.txt rules, which means you can allow or block them selectively.

Some crawlers are used for model training (learning from your content to build the AI), whilst others are used for real-time retrieval (fetching your content when a user asks a relevant question). For AEO purposes, it’s the retrieval crawlers you most want to allow access.

What to do about it

If you’re blocking AI crawlers, intentionally or not, you’re invisible to AI search.

Many sites are blocking them without realising it, often through default CMS settings, security plugins, or overzealous firewall configurations. A bot crawl audit is the fastest way to check.

As of mid-2025, roughly 21% of the top 1,000 websites had specific rules for AI bots in their robots.txt files. This is an active area of website management that most businesses haven’t addressed yet.

Related terms: Bot Crawl Audit, robots.txt, Log File Analysis

AI Overviews: What Are Google’s AI-Generated Answer Boxes?

The short version

Google’s AI-generated answer boxes that appear at the top of search results for certain queries.

Previously known as SGE (Search Generative Experience) during testing, they’re now a core part of Google Search. They’re the reason Google itself is becoming an answer engine.

How it actually works

When Google determines a query would benefit from a synthesised answer, it generates one using its AI models, pulling information from multiple web sources. The AI Overview appears above the traditional blue links, providing a direct answer with expandable source citations.

Not every search triggers an AI Overview. They’re more common for informational and research-style queries than for navigational or simple factual ones. When they do appear, they fundamentally change user behaviour. Users often get what they need from the overview without scrolling down to the traditional results.

The practical takeaway

When AI Overviews appear, click-through rates drop from roughly 15% to 8%, a 46.7% reduction. Only 12% of AI Overviews link to the page that ranks #1 in traditional results. This breaks the long-standing correlation between ranking position and traffic.

The important nuance is that AI Overviews don’t replace SEO. They add a new layer on top. You still need to rank well in traditional results for queries where AI Overviews don’t appear. But for queries where they do, you need your content structured and positioned to be cited as a source in that overview.

This is where AEO and traditional SEO intersect.

Related terms: AEO, Zero-Click Searches, Structural Clarity

AI Referral Traffic: What Is It and How Do You Track It?

The basics

Website visits that come from AI platforms: users who click through to your site from an AI-generated answer in ChatGPT, Perplexity, Google AI Overviews, or other AI search tools.

Under the hood

When an AI system cites your content and includes a link, some users click through. That click shows up in your analytics as referral traffic from the AI platform.

Perplexity is currently the easiest to track. It shows up clearly in Google Analytics referral reports. Google AI Overview clicks appear within your normal Google organic traffic, making them harder to isolate. ChatGPT referral traffic is emerging but volumes are still small.

As of early 2026, AI referral traffic typically sits at around 1% of total traffic for most businesses. The number is small, but it’s growing, and the visitors tend to be more qualified. They’ve already received a detailed answer and are clicking through for a specific reason, which often means they’re further down the purchase funnel.

What to do about it

Set up referral tracking for AI sources in your analytics platform now, even if the volumes are tiny. You want the baseline data so you can track the trend over time.

In Google Analytics 4, check your traffic acquisition report and filter for referral sources including perplexity.ai, chat.openai.com, and claude.ai. Some analytics platforms are starting to add AI traffic as a dedicated channel.

Don’t judge AEO success purely by this metric. Citation and brand visibility in AI answers matter even when users don’t click through. But AI referral traffic is the most directly measurable signal you have right now.

Related terms: Citation, Zero-Click Searches, Brand Sentiment

Bot Crawl Audit: What Is It and Why Does It Matter?

What it is

The process of checking your server logs to see which AI crawlers are visiting your site, how often, and what content they’re accessing (or whether they’re being blocked entirely).

The mechanics

Every time a bot visits your website, it leaves a signature in your server logs through its user-agent string. A bot crawl audit extends traditional Googlebot monitoring to AI-specific crawlers: GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, and others.

The audit checks three things. First, whether your robots.txt file is blocking AI crawlers (many sites are blocking them without realising it). Second, how frequently AI bots are crawling your site and which pages they’re accessing. Third, whether the content they’re finding is structured in a way that’s useful for AI citation.

Why it matters

This is arguably the single most practical starting point for any AEO strategy, and most businesses aren’t doing it.

If AI crawlers can’t access your site, nothing else you do for AEO matters. You’re invisible to the systems generating the answers.

Ask your web team to check two things: your robots.txt file for any rules blocking AI crawlers, and your server logs for AI bot activity. If you’re using a CDN like Cloudflare, check the CDN logs as well.

This takes minutes, costs nothing, and gives you a clear baseline.

Related terms: robots.txt, AI Crawlers, Log File Analysis

Brand Mentions Monitoring: How Do You Track What AI Says About You?

What it is

The practice of regularly checking what AI systems are saying about your business, and tracking how that changes over time. If entity association and cross-web consensus describe how AI forms opinions about your brand, brand mentions monitoring is how you find out what those opinions actually are.

This is sometimes called an entity audit when done as a one-off baseline assessment.

How it works

The simplest version takes about fifteen minutes. Run three types of searches across ChatGPT, Perplexity, and Google’s AI Overviews.

First, search your brand name directly. What does the AI say you do? Is it accurate? Is anything missing or wrong? Second, search your brand name alongside your key topics or services (e.g., “YourBrand commercial equipment hire” or “YourBrand accounting Auckland”). Does the AI associate you with the right things? Third, search the questions your customers actually ask, without mentioning your brand (e.g., “who’s good for commercial equipment hire in the Waikato”). Do you appear in the answer? If not, who does, and why might the AI prefer them?

What comes back gives you a clear picture of your current entity profile: how well AI systems understand who you are, what you do, and whether they trust you enough to recommend you.

For ongoing monitoring, repeat this monthly. The landscape shifts as AI models update, as new content appears across the web, and as your own marketing activity builds (or erodes) your entity associations. A monthly check gives you a trend line rather than a single snapshot.

What it means for you

This is the closest thing to a practical measurement tool that AEO currently has. There’s no equivalent of keyword rankings or traffic dashboards yet. The tools are immature and the data is inconsistent. But a structured entity audit gives you a baseline, and repeating it monthly shows you whether your broader AEO work is moving the needle.

One important caveat: don’t read too much into a single response. Research from SparkToro found that AI systems return a different brand list virtually every time you ask the same question, with less than a 1-in-100 chance of getting the same list twice. A single check tells you whether the AI knows you exist, but to understand your true visibility percentage, you’d need 60 to 100 runs of the same prompt. For most marketing managers, that level of rigour isn’t practical on a manual basis, which is why tracking the trend across monthly audits matters more than obsessing over any individual result.

It also surfaces problems you might not know about. AI might be associating your brand with an outdated service you no longer offer, or confusing you with a similarly named business, or simply not mentioning you at all for queries where you should be appearing. You can’t fix what you don’t know about.

Start with the fifteen-minute version this week. If you want to formalise it, document your results in a simple spreadsheet: date, platform, query, what came back. Over three to six months, the patterns become genuinely useful for guiding your content and PR strategy.

Related terms: Entity Association, Entity Clarity, Brand Sentiment, Confidence Threshold, Consideration Set, Platform Fragmentation

Brand Sentiment: How Do AI Systems Assess It?

The basics

How positively or negatively AI systems perceive and present your brand when generating answers.

It’s your brand’s reputation as understood by AI. Shaped not by what you say about yourself, but by what everyone else says about you across the web.

Under the hood

AI systems build their understanding of your brand by scanning reviews, news articles, forum discussions, social media mentions, and industry publications. They synthesise all of this into a general sentiment profile.

When a user asks “what’s the best [product/service] in [category],” the AI’s recommendation is influenced by this aggregated sentiment.

Unlike traditional reputation management where you can somewhat control the narrative through your own website and paid media, AI sentiment is almost entirely driven by third-party sources. A strong review profile, positive industry mentions, and genuine community presence all feed into how AI systems talk about your brand.

What to do about it

Brand sentiment is becoming a board-level metric alongside traditional performance metrics.

This means investing in the things that build genuine third-party validation: delivering great service that generates positive reviews, earning media coverage through digital PR, being active in industry communities, and publishing original research that others cite.

It’s a longer game than paid advertising, but it’s increasingly what determines whether AI recommends you or your competitor.

Related terms: Entity Association, Third-Party Validation, Cross-Web Consensus, Digital PR

Confidence Threshold: What Determines Whether AI Cites You?

The basics

The level of certainty an AI system needs before it’s willing to include information (or recommend a business) in its generated answer.

If the AI isn’t confident enough in the accuracy of something, it simply leaves it out.

The mechanics

AI systems are designed to sound authoritative. They don’t hedge with “maybe” or “I’m not sure” the way a person might. This means they have internal mechanisms that filter out information that doesn’t meet a certain threshold of confidence.

That threshold is influenced by several factors: how many independent sources confirm the information (cross-web consensus), how consistent the signals are across different platforms (entity clarity), how specific and verifiable the claims are (factual density), and whether trusted sources validate them (third-party validation).

The more of these signals your content and brand provide, the higher the AI’s confidence, and the more likely you are to be cited. If the signals are weak, mixed, or contradictory, the AI’s confidence stays below the threshold and it either omits you entirely or cites a competitor whose signals are cleaner.

Why it matters

This explains why some businesses appear in AI answers and others don’t, even when they seem equally qualified. The business that gets cited isn’t necessarily better. It’s the one the AI can more confidently describe.

It also explains why inconsistency is so damaging. One contradictory signal (an outdated directory listing, a mismatched GBP category, conflicting information across platforms) can drop the AI’s confidence below its threshold. You’re not penalised in the traditional sense. You just get quietly skipped.

The practical implication is that cleaning up inconsistencies and building independent validation isn’t nice-to-have work. It’s directly raising your probability of being cited.

Related terms: Cross-Web Consensus, Entity Clarity, Third-Party Validation, Hallucination

Citation: What Does It Mean in AI Search?

In plain English

When an AI system references your content or brand as a source in its generated answer.

It’s the AI equivalent of being quoted in an article. The system is saying “this information came from here” or “this brand is relevant to your question.”

How it works

Different AI platforms handle citations differently. Perplexity shows numbered source links alongside its answers. Google AI Overviews show expandable source cards. ChatGPT sometimes includes links, sometimes just incorporates information without explicit attribution.

In all cases, the AI has selected your content from among many possible sources because it met the system’s criteria for relevance, trustworthiness, and clarity.

Being cited is distinct from being linked to. A citation means the AI used your content to inform its answer. A link means the user can click through to your site. You can be cited without being linked, which still builds brand visibility and authority even without generating direct traffic.

The practical takeaway

Citation is the new currency of AI search visibility. It’s the AEO equivalent of ranking on page one.

Your goal is to become a source that AI systems consistently turn to when answering questions in your area of expertise.

To increase your chances of being cited, focus on creating content that provides clear, specific, evidence-backed answers (factual density), is structured so AI can extract discrete pieces of information (structural clarity), and is validated by being referenced across multiple trusted platforms (cross-web consensus).

Related terms: AEO, Factual Density, Structural Clarity, Cross-Web Consensus

Citation Recency: Why AI Favours Recent Coverage

What it is

The strong tendency of AI systems to cite recently published content over older material when generating answers. Research analysing over one million AI citations found that half came from content published within the previous 11 months, with the single most common citation day being yesterday.

This isn't the same as content freshness (whether your own content is up to date). Citation recency is about how recently others have written about you or your area of expertise.

How it works

AI systems using real-time retrieval (RAG) actively search the current web when generating answers. They weight recent sources more heavily, particularly for queries that imply a need for current information. Coverage from six months ago carries less weight than coverage from last month, and coverage from last week carries more weight than both.

The recency weighting varies by AI platform. ChatGPT weights recent journalism more aggressively, with over half of its cited journalism links published within the past year. Claude is less aggressive, with around 36% of cited journalism from the last year. But the pattern holds across all platforms: recent coverage gets cited more.

This creates a visibility half-life. Your AI presence decays over time unless replenished by new third-party mentions and coverage. A burst of media activity that generates ten pieces of coverage in January followed by silence until September means you're effectively less visible to AI systems for most of the year.

What it means for you

Citation recency transforms how you need to think about media activity and third-party presence. The traditional model of periodic PR campaigns, generating a wave of coverage around a launch then moving on, doesn't sustain AI visibility the way it sustains traditional search rankings.

What works is cadence: a steady drumbeat of reasons for others to talk about you. That doesn't mean more work, it means differently distributed work. Instead of concentrating all your media activity around announcements, spread it across the year.

For most NZ businesses, achievable cadence might look like one piece of contributed expertise per quarter to a trade publication, regular commentary on industry developments, and consistent activity on platforms where your industry conversations happen. The shift in mindset is from "what do we want to announce" to "what do we know that others would find valuable."

Related terms: Content Freshness, Digital PR, Earned Media Signals, Visibility Decay, Citation

Consideration Set: Why AI Doesn't Rank You, It Decides Whether to Include You

What it is

The stable pool of brands that an AI system draws from when answering a particular type of question. Unlike traditional search rankings, where businesses occupy fixed positions on a list, AI systems maintain a consideration set of brands they might recommend, then select from that pool differently each time.

You're either in the consideration set or you're not. If you're in it, you'll appear in some percentage of AI responses. If you're not, you won't appear at all, no matter how many times someone asks.

How it works

Research from SparkToro analysing 2,961 prompts across AI platforms found that there is less than a 1-in-100 chance any AI tool returns the same brand list twice for the same prompt. The same list in the same order appears less than 0.1% of the time. Every response is different.

But here's the important part: despite that apparent randomness, the top brands in any category appear in 55 to 77% of responses. The individual lists are always different, but the pool of brands being drawn from is remarkably stable. This is the consideration set.

AI systems don't generate a fixed ranking and read from it. They make a probabilistic selection from a set of entities they're confident enough to recommend. Your visibility percentage across many responses is far more stable and meaningful than your position in any single response.

What determines whether you're in the consideration set is the same cluster of signals covered throughout this glossary: entity association, cross-web consensus, entity consistency, and third-party validation. These build the AI's confidence in your brand to the point where you become part of the pool it draws from.

What it means for you

This fundamentally changes how you should think about AI search measurement. Any tool showing you a single "AI ranking position" is unreliable, because that position changes with every prompt. Research suggests you need 60 to 100 prompt runs to get meaningful visibility data.

The question isn't "what position are we?" It's "are we in the consideration set at all, and what percentage of responses include us?" A brand appearing in 60% of AI responses for its category queries is performing well, even though it never holds a fixed "position."

For marketing managers, this means tracking trends over time rather than chasing specific positions. Monthly entity audits that check whether you appear in AI responses for your key queries, tracked over three to six months, will show you whether your consideration set presence is growing, stable, or declining. That trend line is far more useful than any snapshot ranking.

It also means that the competitive threshold matters. In smaller or less competitive categories, the bar for entering the consideration set is lower. Fewer brands competing for the same entity associations means you need less total signal to be included.

Related terms: Confidence Threshold, Entity Association, Brand Mentions Monitoring, Cross-Web Consensus, Entity Consistency

Content Freshness: Why Outdated Content Hurts Your AI Visibility

What it is

How recently your content has been published or meaningfully updated. AI systems, particularly those using real-time retrieval, favour current content over outdated material when generating answers.

This isn’t about publishing dates for the sake of it. It’s about whether your content reflects the current state of your industry, your services, and your expertise.

How it works

AI search operates across two layers. The training layer absorbs your accumulated web presence over months and years. The retrieval layer searches the current web in real time when a user asks a question.

For the retrieval layer, content freshness matters directly. When an AI system pulls in sources to build an answer, it factors in how current those sources are. A well-structured, factually dense article from last month will typically be preferred over an equally good article from three years ago, particularly for topics where information changes.

This doesn’t mean old content is worthless. Evergreen content that’s still accurate retains value, especially for the training layer where your historical web presence contributes to entity associations. But content that’s visibly outdated, referencing old statistics, discontinued products, or superseded information, actively works against you. It signals to AI systems that your information may not be reliable.

The risk is compounding. If a significant portion of your website hasn’t been updated in years, AI systems may reduce their confidence in your content overall, not just in the outdated pages.

What it means for you

Most businesses have a content decay problem they haven’t addressed. Blog posts from 2019 with outdated statistics. Service pages that still reference old pricing or discontinued offerings. Case studies featuring clients who’ve moved on.

The practical approach is straightforward. Audit your existing content and sort it into three categories: still accurate (leave it), needs updating (refresh the data and republish with a current date), and no longer relevant (consider removing or consolidating).

Prioritise the pages that matter most for your AEO visibility: your core service pages, your most authoritative content, and anything that ranks well or gets cited. You don’t need to update everything at once. Focus on the content that AI systems are most likely to retrieve.

Building a regular content review into your calendar, even quarterly, prevents the gradual staleness that weakens your retrieval-layer visibility over time.

Related terms: RAG, Factual Density, Structural Clarity, Information Gain

Cross-Web Consensus: What Is It and Why Does It Matter?

The short version

The principle that AI systems give more weight to information that’s confirmed across multiple independent, trustworthy sources.

If several unrelated websites agree on something, the AI treats it as more reliable than a single source claiming it.

How it actually works

When an AI system researches a topic to generate an answer, it doesn’t just take the first result it finds. It cross-references information across multiple sources, looking for agreement.

If your brand’s expertise in a topic is mentioned on your website, in an industry publication, on a review platform, and in a forum discussion, that convergence of independent signals gives the AI much higher confidence than any single source alone.

This is conceptually similar to how academic citation works. A finding is more credible when it’s been independently replicated. AI systems apply the same logic at web scale.

What it means for you

Multi-platform distribution is essential for AEO. Content that only exists on your website has a single point of validation. Content that’s been picked up, referenced, discussed, and validated across multiple platforms has the cross-web consensus that AI systems look for.

This doesn’t mean duplicating your content everywhere. It means creating content worth referencing: original research, unique data, distinctive perspectives. Then actively distributing and promoting it so it gets picked up by industry publications, discussed in communities, and cited by others.

Generic content that says what everyone else says won’t build consensus because it doesn’t add anything new to reference.

Related terms: Third-Party Validation, Entity Association, Confidence Threshold, Factual Density

Digital PR: Why Earned Media Now Drives AI Visibility

What it is

The practice of earning media coverage, industry mentions, and expert citations across third-party platforms. In the AEO context, digital PR serves a specific strategic purpose: building the entity associations and cross-web consensus that AI systems use to determine whether to recommend your business.

Traditional PR measured success in reach and impressions. Digital PR for AEO measures success in whether your brand appears alongside the right topics on platforms AI systems trust.

How it works

When an industry publication quotes your expert on a topic, that creates an entity association between your brand and that subject. When a news site mentions your business in the context of your industry, that contributes to cross-web consensus. When a trade journal features your case study, that builds third-party validation. Each of these signals independently strengthens your entity profile.

The critical difference from traditional PR is that links are no longer the primary currency. An unlinked mention of your brand in a respected industry article carries significant weight for AI systems, even though traditional SEO would give it almost no credit. This means the value of PR activity is broader than it used to be. A podcast appearance, a conference write-up, a contributed expert comment: all build your entity profile regardless of whether they include a link back to your website.

The platforms that matter most are the ones AI systems already trust and frequently draw from: established industry publications, major news outlets, professional directories, and active community forums. Getting mentioned once on a high-authority platform in the right context can be worth more than dozens of mentions on low-quality sites.

What it means for you

If you’re not investing in digital PR, you’re missing one of the most effective levers for AEO visibility. It’s how you build the third-party validation and cross-web consensus that move you above the confidence threshold.

The practical starting points are contributing expert commentary to industry publications, pitching original research or data to journalists, speaking at industry events (which generates write-ups and mentions), and building relationships with the publications your audience reads.

This isn’t quick-win work. Building genuine media relationships and earning coverage takes time and consistency. But the signals compound. Each mention strengthens your entity associations, and the cumulative effect across multiple platforms is what pushes you past the confidence threshold for AI citation.

For marketing managers working with limited budgets, focus on your niche. Being consistently mentioned in two or three respected industry publications carries more weight than scattered coverage across dozens of generic sites.

Related terms: Entity Association, Unlinked Mentions, Third-Party Validation, Cross-Web Consensus, Brand Sentiment

Earned Media Signals: Why Independent Coverage Is the Primary Driver of AI Citation

What it is

The credibility signals that independent media coverage sends to AI systems. When a journalist, trade publication, industry body, or independent expert writes about your business without being paid to do so, AI systems treat that coverage as a stronger trust signal than anything you publish yourself.

Research analysing over one million AI citations across ChatGPT, Claude, and Gemini found that 94% of all citations came from non-paid sources, with 82% from earned media alone. Your own website, your paid placements, and your sponsored content account for a small fraction of what AI systems actually cite.

How it works

AI systems face a fundamental credibility challenge: they need to sound authoritative without being able to personally verify what's true. Their solution is to lean heavily on independent sources, because independence functions as a proxy for reliability. Your website can never be independent of you. A journalist writing about your industry can be.

This isn't a preference or a weighting factor. It's structural. AI systems are designed to minimise the risk of recommending something inaccurate or biased. Independent coverage provides both consensus (multiple sources agreeing) and independence (sources with no financial incentive to promote you) simultaneously. That combination is what pushes entities past the confidence threshold for citation.

The types of earned media that carry the most weight are coverage from established industry publications, expert commentary cited in news reporting, independent reviews and assessments, conference coverage and event write-ups, and industry body publications. What matters is that the source has editorial independence, the coverage isn't something you paid for or directly controlled.

What it means for you

For most NZ businesses, this is genuinely encouraging rather than daunting. The companies that invested heavily in paid PR were largely targeting the wrong outcomes for AI visibility anyway. Research shows only a 2% overlap between the journalists that PR teams typically pitch and the sources AI systems actually cite. The game isn't about having a big PR budget. It's about being genuinely useful to the publications and platforms that AI systems trust.

The practical shift is from "what do we want to announce" to "what do we know that others would find valuable." One meaningful piece of contributed expertise per quarter to a relevant trade publication does more for your AI visibility than a dozen paid placements. In a smaller market like New Zealand, the bar for becoming a recognised voice in your category is lower than in larger markets, and trade publications and industry bodies carry disproportionate weight.

Related terms: Digital PR, Third-Party Validation, Confidence Threshold, Cross-Web Consensus, Citation Recency, Pitch Gap

Entities (in Search and AI Context)

What it is

A distinct, identifiable concept that search engines and AI systems recognise independently of the words used to describe it. Your business is an entity. So is each of your products, your locations, your industry, and the topics you want to be known for. When we talk about entity association, entity consistency, and cross-web consensus elsewhere in this glossary, this is the foundational concept they all build on.

How it works

Traditional search engines matched words. If someone searched “best accountant Auckland,” Google looked for pages containing those words and ranked them based on links, domain authority, and on-page optimisation. The search engine was essentially pattern-matching strings of text.

Modern search engines, and AI systems in particular, work differently. They maintain vast databases of entities (Google’s is called the Knowledge Graph) where every recognisable concept has its own unique identifier. Your business isn’t just a name on a web page. It’s a record in a database, connected to other records: the industry you operate in, the services you provide, the locations you serve, the people who work there.

When someone asks an AI system a question, it doesn’t just search for matching words. It identifies the entities in the question, then looks for content and sources that are strongly associated with those entities. If someone asks “who’s good for commercial equipment hire in the Waikato,” the AI identifies three entities (commercial equipment hire, Waikato, and the implicit concept of a service provider) then looks for businesses whose entity profiles are strongly connected to all three.

This is why two businesses can have almost identical website content, but one gets cited by AI and the other doesn’t. The difference isn’t the words on the page. It’s how well connected their entity is to the right topics across the web.

What it means for you

Understanding entities explains why so many of the strategies in this glossary work the way they do. Entity association matters because AI systems recommend businesses whose entity profiles are strongly connected to relevant topics. Cross-web consensus matters because multiple independent sources mentioning your brand alongside a topic strengthens that entity connection. Entity consistency matters because inconsistent information fragments your entity profile and weakens AI confidence.

The practical shift is this: you’re not just optimising web pages anymore, you’re building and strengthening your business’s entity profile across the entire web. Every review, every industry mention, every directory listing, every piece of structured data on your website contributes to how AI systems understand what your business is and what it’s connected to.

Start by searching for your business name in ChatGPT, Perplexity, and Google’s AI Overviews. What comes back tells you how well-defined your entity profile currently is. If the AI knows who you are, what you do, and where you operate, you have a strong foundation. If it’s vague, confused, or wrong, that’s your starting point for improvement.

Related terms: Entity Association, Entity Consistency, Entity Salience, Cross-Web Consensus, Schema Markup, Google Business Profile

Entity Association: How Does It Work?

What it is

The way AI systems map relationships between your brand and specific topics based on how often they appear together across the web.

Think of it as your brand’s digital reputation. Not based on links, but on context and co-occurrence.

The mechanics

Traditional search authority worked like a voting system. If a high-authority site linked to you, they handed you a vote of power. You could rank without anyone actually mentioning your brand name, as long as enough sites pointed links at your domain.

AI entity association works differently. Instead of counting links, AI systems scan the web to see how closely your brand name sits next to specific topics in sentences and paragraphs.

If a trusted industry report mentions your company as a leader in a specific area but doesn’t link to you, traditional SEO would give you almost zero credit. For an AI system, that unlinked mention is a significant signal.

The AI creates a mathematical relationship between your brand and that subject based on co-occurrence in reputable “neighbourhoods” of the internet: industry journals, news sites, major forums, and review platforms.

Why it matters

Your marketing strategy needs to incorporate building conversations, not just link pipelines. You need your brand name appearing in the same context as your target topics on platforms AI already trusts.

Practical tactics include digital PR (getting quoted in industry publications), publishing original research that others reference, building genuine community presence in forums and discussion platforms, and earning reviews on relevant platforms.

Being talked about in the right context is now as valuable as being linked to.

Related terms: Unlinked Mentions, Third-Party Validation, Cross-Web Consensus, Trust Cascade, Digital PR

Entity Clarity: What Is It and Why Does It Determine Visibility?

The basics

Whether AI systems can confidently identify who you are, what you do, and what you’re known for.

Under the hood

When an AI system is pulling together an answer to someone’s question, it’s drawing on everything it’s encountered about your business across the web: your website, your Google Business Profile, industry directories, review sites, media mentions, social profiles.

If all of those sources tell a consistent, clear story, the AI has high confidence in recommending you. If the signals are messy or contradictory, it moves on to someone else.

The plain language version: entity clarity is your business’s digital consistency. Same name, same description of what you do, same core messaging, everywhere you appear online. Not identical copy on every platform, but a coherent story that all points in the same direction.

What to do about it

AI systems don’t take risks with their answers. They’re designed to sound authoritative, so they gravitate towards businesses they can confidently describe.

If your signals are muddled (maybe your website says one thing, your LinkedIn says another, your directory listings are outdated, and your Google Business Profile categories don’t quite match what you actually do) the AI just picks someone cleaner.

You don’t get argued with or penalised. You just get quietly skipped.

The simplest audit: Google your own business name. Look at what comes back. Does your Google Business Profile match what your website says? Do the directory listings on page one have the right description, the right categories, the right phone number? If there’s a Knowledge Panel, is the information accurate?

Three searches. Ten minutes. That’s your baseline.

Related terms: Entity Association, Entity Consistency, Google Business Profile, Third-Party Validation

Entity Consistency: The Foundation AI Systems Check First

What it is

Whether the key facts about your business, your name, what you do, where you operate, how to contact you, are the same everywhere you appear online. AI systems cross-reference information about your business across multiple sources. When the details match, confidence goes up. When they don’t, confidence goes down.

Entity consistency is closely related to entity clarity. Clarity is about whether AI systems can confidently identify your business. Consistency is the specific mechanism that either builds or undermines that clarity.

How it works

Every time your business appears online, whether on your website, Google Business Profile, an industry directory, a review platform, or a social media profile, AI systems treat that as a data point about your entity. They compare these data points against each other.

If your business name is slightly different across platforms (say, “Smith & Associates” on your website, “Smith and Associates Ltd” in a directory, and “Smith Associates” on LinkedIn), AI systems have to decide whether these are the same entity or different ones. The more variation, the harder that decision becomes, and the less confident the AI is in any single version.

The same applies to categories, descriptions, addresses, phone numbers, and service lists. AI systems aren’t just checking whether you exist. They’re checking whether the information about you is reliable. Contradictory signals don’t average out to a moderate level of confidence. They actively reduce confidence because the AI can’t determine which version is correct.

For multi-location businesses, this multiplies. If you have thirty locations and each one has slightly different information across a dozen platforms, that’s hundreds of potential inconsistencies fragmenting your entity profile.

What it means for you

The practical priority is straightforward: audit your business information across every platform where you appear and make it consistent.

Start with Google Business Profile, since that’s the most influential structured data source for AI systems. Make sure your business name, categories, description, address, phone number, and services are exactly right. Not approximately right. Exactly right.

Then work outward to the platforms that appear when you search for your business name: major directories, review sites, LinkedIn, industry listings. Fix any discrepancies. Update anything that’s outdated.

This isn’t a one-off task. Information drifts over time as platforms update their own records, as your business evolves, or as old listings get out of date. Building a regular consistency check into your marketing calendar, even quarterly, prevents the gradual fragmentation that weakens your entity profile.

The good news is that this work compounds. Every inconsistency you fix directly strengthens the AI’s confidence in recommending you.

Related terms: Entity Clarity, Entities, Entity Association, Google Business Profile, Confidence Threshold

Entity Salience: Why Not All Mentions Are Created Equal

What it is

A measure of how central your business is to a piece of content, not just whether it’s mentioned. AI systems don’t treat all mentions equally. Being the subject of an in-depth case study sends a far stronger signal than appearing as one name in a list of fifty providers. Entity salience is the difference between being talked about and being talked about in depth.

How it works

When AI systems process a web page, they don’t just note which entities appear. They assess how prominent each entity is within the content. A page that dedicates three paragraphs to explaining how your business solved a specific problem gives your entity high salience on that page. A directory listing that includes your name alongside hundreds of others gives your entity very low salience.

Several factors influence salience. Position matters: entities mentioned in headings, opening paragraphs, and page titles carry more weight than those buried in a footnote. Depth matters: content that describes what your business does, how it operates, and what makes it distinct gives AI systems richer entity information than a bare name and link. Context matters: being mentioned alongside closely related entities (your industry, your location, your specific services) reinforces the right associations more strongly than appearing in unrelated content.

This is one reason why manufacturing mentions at scale tends to backfire. A hundred low-salience mentions across generic directories contribute far less to your entity profile than ten pieces of content where your business is genuinely central to the discussion.

What it means for you

When evaluating where your business gets mentioned across the web, think about quality of presence, not just quantity. An industry publication writing about your approach to a specific challenge is worth more for your entity profile than dozens of directory listings where you’re one name among many.

This has practical implications for where you focus effort. Contributing expert commentary to an industry article, being featured in a case study, or having a journalist write about your work all produce high-salience mentions that strengthen your entity profile significantly. Bulk submissions to generic directories produce low-salience mentions that add very little.

It also applies to your own content. Pages on your website that go deep on specific topics, explaining your expertise and approach in detail, give AI systems strong salience signals about what your business is known for. Thin pages that mention a dozen services without depth give the AI very little to work with.

The practical question to ask: when AI systems encounter mentions of my business across the web, are we central to the conversation or just part of the background noise?

Related terms: Entities, Entity Association, Cross-Web Consensus, Confidence Threshold, Information Gain

Trust Cascade: How Does AI Build Confidence in Your Brand?

In plain English

The cumulative effect of multiple trust signals reinforcing each other across platforms, progressively building AI systems’ confidence in your brand until you cross the threshold for citation.

No single signal gets you cited. It’s the combination.

Under the hood

Think of it like a cascade of evidence. Your website states your expertise clearly (entity clarity). An industry publication mentions you in the same context (entity association). Google reviews confirm your quality (third-party validation). A forum discussion recommends you (cross-web consensus). Your Google Business Profile matches everything else (consistency).

Each signal individually is modest. Together, they create a compounding effect. Each one raises the AI’s confidence, and the combination pushes you over the citation threshold.

The cascade works in reverse too. If most of your signals are strong but one platform contradicts the others (outdated directory listing, incorrect GBP categories, a negative review trend) that weakens the entire cascade. The AI doesn’t average out your signals. One weak link creates doubt that affects confidence across the board.

What to do about it

Map your trust signals across platforms. Website, Google Business Profile, directories, review platforms, industry publications, social profiles, forum presence. Are they all telling the same story? Are there gaps or contradictions?

Then prioritise building the signals you’re missing. If you have strong website content but no third-party validation, focus on PR and reviews. If you have great reviews but weak entity clarity, clean up your directory listings and GBP.

The goal isn’t perfection on one platform. It’s consistency across all of them.

Related terms: Confidence Threshold, Entity Clarity, Cross-Web Consensus, Third-Party Validation

Factual Density: What Is It and Why Does It Matter?

In plain English

The concentration of specific, verifiable facts and data points in your content, as opposed to vague claims, filler, or general statements.

AI systems prefer content that packs a lot of concrete information into a clear structure.

How it works

When an AI system evaluates potential sources to cite, it’s looking for content that provides specific answers, not waffle.

A page that says “our product is really good and lots of customers love it” gives the AI nothing to work with. A page that says “94% of customers rated our service 4+ stars, with average resolution time of 2.3 hours” gives it citable facts.

Factual density doesn’t mean cramming every sentence with numbers. It means ensuring your content contains specific, substantiated claims rather than vague assertions. Original data, concrete examples, named sources, and verifiable statistics all contribute to factual density.

The practical takeaway

Review your key content through this lens: if an AI system needed to extract three facts from this page, could it?

If your service pages, about page, and key articles are full of generic marketing language, they’re unlikely to be cited regardless of how well they rank.

The most effective approach is to incorporate original data wherever possible. Customer statistics, survey results, case study outcomes, industry benchmarks: these are the kinds of specific facts that AI systems cite and that generic AI-generated content can’t replicate.

Related terms: Structural Clarity, Citation, Information Gain

Google Business Profile: Why Does It Matter for AI Search?

What it is

Google’s free business listing tool (formerly Google My Business) that controls how your business appears in Google Search and Google Maps.

For AI search, your Google Business Profile is one of the strongest entity signals available. It directly tells Google (and indirectly tells other AI systems) who you are, what you do, and where you operate.

The mechanics

Your Google Business Profile contains structured information about your business: name, address, categories, description, hours, services, photos, reviews, and Q&A. Google uses this data not just for Maps and local search, but as a foundational entity signal for AI Overviews.

When someone asks an AI system a local query (“best accountant in Auckland” or “gym franchises near me”) the information in Google Business Profiles feeds directly into the answer. The categories you’ve selected, the reviews you’ve accumulated, and the consistency between your GBP and the rest of your web presence all influence whether you get cited.

For multi-location businesses, each location’s GBP acts as a separate entity signal. Inconsistencies between locations (different category selections, outdated descriptions, missing attributes) weaken entity clarity across the entire brand.

Why it matters

If you’re already maintaining your Google Business Profile for local search, you’re most of the way there for AI search. The additional step is making sure your GBP information is precisely consistent with what the rest of the web says about you.

Check that your primary and secondary categories accurately reflect what you do. Make sure your business description matches your website’s positioning. Ensure your attributes, services, and Q&A are up to date. For multi-location businesses, audit consistency across all locations.

A well-maintained GBP is one of the few things you fully control that directly feeds into AI systems’ understanding of your business.

Related terms: Entity Clarity, Entity Consistency, Entity Association, Schema Markup

GEO (Generative Engine Optimisation): What Is It and How Does It Differ from AEO?

The short version

Optimising your content specifically for generative AI search engines: platforms like ChatGPT, Perplexity, and Google’s AI features that generate answers rather than just listing links.

GEO is closely related to AEO and the terms are often used interchangeably, though GEO specifically emphasises the generative AI aspect.

How it actually works

GEO as a term emerged from academic research and has been adopted by parts of the industry to describe optimisation specifically for AI-generated search results. The underlying principles are the same as AEO: creating content that AI systems can easily find, trust, extract from, and cite.

Some practitioners draw a distinction where AEO is the broader concept (optimising for any answer engine, including featured snippets and knowledge panels) whilst GEO is specifically about generative AI systems. In practice, the strategies overlap almost entirely.

What it means for you

Don’t get hung up on the terminology. Whether someone calls it AEO or GEO, they’re talking about the same fundamental shift: optimising for citation in AI-generated answers rather than just ranking in traditional results.

The practical strategies (structural clarity, entity association, cross-web consensus, factual density) apply regardless of which label you use.

Related terms: AEO, AI Overviews, RAG, Citation

Hallucination: What Happens When AI Gets Your Business Wrong

What it is

When an AI system generates information that sounds confident and plausible but is factually incorrect. In a business context, this might mean AI telling a user you offer services you don’t, attributing quotes to you that you never said, confusing your business with a similarly named one, or stating incorrect details about your locations, pricing, or history.

How it works

AI systems are designed to produce fluent, confident-sounding responses. When they don’t have enough reliable information about a topic, they don’t say “I don’t know.” They fill the gap with their best guess based on patterns in their training data, and that guess can be completely wrong while sounding entirely authoritative.

Hallucination is more likely when your entity profile is thin (not enough information about your business across the web for the AI to draw on), when signals are contradictory (the AI has conflicting information and resolves it incorrectly), or when your brand name is similar to another business (the AI merges entity profiles).

This is the flip side of the confidence threshold concept. When AI has strong, consistent signals about your business, it generates accurate responses. When the signals are weak or mixed, it either skips you entirely (which is frustrating but not harmful) or fills in the blanks with fabricated details (which can actively damage your reputation).

What it means for you

If AI is saying something wrong about your business, that’s not just an abstract AI problem. It’s a signal that your entity profile needs work.

Start with the entity audit: search your business name across ChatGPT, Perplexity, and Google’s AI Overviews and note anything inaccurate. Common issues include outdated information (old addresses, discontinued services, former staff listed as current), conflated identities (your business confused with a similarly named one), and fabricated details (services you don’t offer, locations you don’t have, awards you haven’t won).

The fix isn’t to contact AI companies and ask them to correct it. The fix is to strengthen the underlying signals. Clean up entity consistency across platforms. Ensure your Google Business Profile is precisely accurate. Build more independent mentions that correctly describe what you do. Over time, as these stronger signals accumulate, the AI’s responses become more accurate.

Understanding hallucination also helps manage expectations internally. If your CEO asks “why is ChatGPT saying we do X when we don’t,” you can explain the mechanism and connect it to the broader AEO work you’re doing to strengthen your entity profile.

Related terms: Confidence Threshold, Entity Clarity, Entity Consistency, Brand Mentions Monitoring

Information Gain: What Is It and Why Does Generic Content Fail?

The short version

The amount of new, unique information your content adds beyond what’s already widely available on a topic.

AI systems prioritise content with high information gain because it gives them something worth citing that they can’t get from a dozen other sources.

How it actually works

When an AI system evaluates multiple sources on the same topic, it’s not just looking for accuracy. It’s assessing what each source adds that the others don’t. Content that merely restates what every other page says has zero information gain. Content that adds original data, a unique perspective, proprietary research, or specific experience has high information gain.

This is the mechanism behind why generic AI-generated content is, as one industry report put it, “essentially worthless” for AI visibility. If you use AI to produce the same broad overview that every competitor’s AI also produced, no source stands out. There’s nothing for the answer engine to specifically cite because everything says the same thing.

The practical takeaway

This has direct implications for your content strategy. The content most likely to be cited by AI systems is content that only you can create: your original research, your proprietary data, your specific customer outcomes, your expert perspective based on actual experience.

Before creating content, ask: what can we say about this topic that nobody else can? If the answer is “nothing,” the content might be useful for other purposes but it’s unlikely to earn AI citation.

Information gain is also why thought leadership (genuine thought leadership based on real expertise, not repackaged common knowledge) has become strategically important for AI search visibility.

Related terms: Factual Density, Cross-Web Consensus, Citation

llms.txt: Should You Create One?

What it is

A proposed standard for a text file (similar to robots.txt) that websites can publish to provide AI systems with a summary of their content, structure, and key information in a format specifically designed for large language models.

The idea is that instead of AI crawlers having to figure out what your site is about by reading every page, you give them a concise briefing document.

How it works

The llms.txt file would sit at the root of your website (e.g., yoursite.co.nz/llms.txt) and contain a structured summary of your site: what the business does, what content is available, how it’s organised, and which pages are most important. AI systems could read this file first to get oriented before crawling specific pages.

The concept has generated significant discussion in the SEO and AEO community. Some practitioners have advocated for it as an essential part of AI search strategy.

However, the evidence so far doesn’t support that enthusiasm. Research across 300,000 domains found no correlation between having an llms.txt file and being cited in AI-generated answers. As of early 2026, major AI providers haven’t implemented support for the format, meaning the file exists but AI systems aren’t specifically looking for it or using it.

What it means for you

Right now, llms.txt is not something you need to act on. It’s an interesting proposal that may gain traction in the future, but there’s no evidence it influences AI citation today.

Your time is better spent on things that demonstrably matter: ensuring AI crawlers can access your site through robots.txt, building structural clarity in your content, strengthening entity consistency across platforms, and earning third-party validation.

If llms.txt does gain adoption by major AI providers in the future, this entry will be updated accordingly. For now, file it under “interesting to know about, nothing to do yet.”

Related terms: robots.txt, AI Crawlers, Bot Crawl Audit, Structural Clarity

Log File Analysis: What Is It and How Does It Support AEO?

The basics

The process of examining your web server’s log files to understand which bots are visiting your site, which pages they’re accessing, and how often.

For AEO, it’s how you find out whether AI systems are actually reading your content, or whether they’re being turned away at the door.

The mechanics

Every time any bot visits your website, your server records the visit in a log file. That record includes the bot’s identity (its user-agent string), which page it requested, and whether the request was successful or blocked.

Traditional log file analysis focused on Googlebot: understanding how Google was crawling your site and whether it was finding your important pages. AEO log file analysis extends this to AI-specific crawlers: GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, and others.

The data tells you three things. Whether AI bots are visiting at all (if not, check your robots.txt). Which pages they’re spending time on (this shows what AI systems consider your most valuable content). And how frequently they’re returning (more frequent crawls suggest the AI considers your content worth keeping current).

Why it matters

Most businesses have never looked at their log files for AI crawler activity. It’s the data layer behind a bot crawl audit. The audit tells you whether you’re blocking crawlers. Log file analysis tells you the full picture of how AI systems are interacting with your site over time.

If you’re working with a web team or SEO agency, ask them to filter your server logs for AI bot user-agent strings. If you’re on a managed hosting platform, the logs may be accessible through your control panel. CDN platforms like Cloudflare also provide bot traffic data.

This isn’t something you need to check daily. A monthly review gives you a clear trend line.

Related terms: Bot Crawl Audit, AI Crawlers, robots.txt

Modular Content: What Is It and Why Does AI Prefer It?

In plain English

Content that’s organised in self-contained sections, each of which can be understood and extracted independently without needing the surrounding context.

It’s the difference between a wall of text and a well-organised reference page where each section stands on its own.

How it works

AI systems don’t read your page top to bottom like a human might. They scan for the specific section that answers the user’s question, extract that section, and cite it, often without referencing anything else on the page.

Modular content supports this by ensuring each section has a clear header that signals what it covers, contains a complete answer within that section (not split across multiple parts of the page), and makes sense even if extracted in isolation from the rest of the content.

The opposite of modular content is a long, flowing essay where ideas build on each other and you need to read from the beginning to understand any individual point. That format works for storytelling but fails for AI extraction.

The practical takeaway

When creating or restructuring content, ask yourself: if an AI system pulled just this one section out of context, would it make sense as a standalone answer?

If the answer is no, the section either needs more context added within it, or it needs to be restructured so the key information is self-contained.

This doesn’t mean every page has to be a Q&A format. It means your important pages should be structured so each section delivers a complete, extractable answer, even within longer, more comprehensive content.

Related terms: Structural Clarity, Schema Markup, Factual Density

Pitch Gap: Why Traditional PR Targeting Misses the Sources AI Actually Cites

What it is

The significant misalignment between the journalists and publications that PR teams traditionally target and the sources that AI systems actually draw from when generating answers. Research found only a 2% overlap between the two: 98% of the sources AI cites are not the ones PR professionals have been pitching.

How it works

Traditional PR targets publications based on reach, audience size, and prestige. The logic has always been straightforward: get coverage in the biggest outlets to reach the most people. This produced a well-established playbook focused on major newspapers, popular magazines, broadcast media, and high-traffic websites.

AI citation operates on entirely different criteria. AI systems select sources based on factual density, editorial independence, topical authority, and structural clarity, not audience size. A specialist trade publication with a fraction of the readership of a national newspaper might be cited far more frequently by AI systems for queries in that specialist's domain.

The result is a mismatch. PR teams have spent years building relationships with journalists at major outlets, while the journalists whose work AI systems most frequently cite are often specialists, trade reporters, and niche experts that traditional PR programmes overlook entirely.

What it means for you

If you're already investing in PR, this isn't a reason to stop. It's a reason to redirect. The media relationships that drive AI visibility are not necessarily the ones your PR programme is currently prioritising.

For NZ businesses, this is actually an advantage. In a smaller market, the number of specialist trade publications and industry commentators is manageable. You can realistically map the publications that matter for AI visibility in your category and build relationships with the specific journalists whose work gets cited.

The practical starting point is an audit: which publications appear when you ask AI systems questions about your industry? Those are the sources AI already trusts. Building relationships with those specific outlets and contributors, even if they're smaller than the publications your current PR targets, will have a disproportionate impact on your AI visibility.

Related terms: Earned Media Signals, Digital PR, Third-Party Validation, Entity Association, Citation

Platform Fragmentation: Why Visibility on One AI Platform Doesn't Mean Visibility on All

What it is

The significant variation in which sources different AI platforms cite when answering the same question. ChatGPT, Claude, Gemini, and Perplexity each draw from different source pools, weight different signals, and produce different answers, even for identical queries. Being well-cited on one platform is no guarantee of visibility on another.

How it works

Each AI platform has its own training data, its own retrieval mechanisms, and its own approach to evaluating source credibility. The result is that the same question can produce substantially different answers with different sources cited across different platforms.

Research from Muck Rack found striking differences in how platforms cite even major publishers. Reuters, for example, is cited roughly 20 times less frequently by Claude than by Gemini, and around 50 times less frequently than by ChatGPT. These aren't small variations. They represent fundamentally different source preferences across platforms.

This fragmentation exists because each platform's retrieval system indexes and weights sources differently. A trade publication that Perplexity draws from heavily might barely register in ChatGPT's citation patterns. An industry blog that ChatGPT frequently cites might not appear in Claude's preferred sources at all.

The fragmentation extends beyond just which sources get cited. Different platforms also show different sensitivity to citation recency, different thresholds for confidence before citing, and different patterns for how they handle entity associations. Your brand might be well-represented in one platform's understanding and largely absent from another's.

What it means for you

The practical implication is that you can't optimise for "AI search" as a single channel. You need to check your visibility across multiple platforms and understand that your presence will differ across each one.

When running entity audits, always check across at least ChatGPT, Perplexity, and Google's AI Overviews. If you're only checking one platform, you're getting an incomplete picture. A brand that appears consistently across all three platforms has a genuinely strong entity profile. A brand that appears on one but not the others has platform-specific visibility that could disappear if user behaviour shifts.

The good news is that the underlying fundamentals, entity consistency, cross-web consensus, third-party validation, and earned media signals, strengthen your position across all platforms simultaneously. You don't need a separate strategy for each AI system. You need a strong enough entity presence that all of them pick it up. Platform fragmentation is a measurement reality, not a reason to fragment your strategy.

Related terms: Brand Mentions Monitoring, Earned Media Signals, Citation Recency, Consideration Set, Cross-Web Consensus

RAG (Retrieval-Augmented Generation): What Is It and Why Should Marketers Care?

The short version

The technical process AI systems use to generate answers: they retrieve relevant content from the web first, then use that content to generate their response.

It’s the mechanism that makes AEO possible, and understanding it explains why structural clarity and factual density matter so much.

How it actually works

When you ask an AI search tool a question, two things happen in sequence. First, the system retrieves. It searches the web (or its index of the web) for content relevant to your question, pulling in pages, sections, and data points from multiple sources. Second, it generates. It uses its AI model to synthesise those retrieved sources into a coherent answer.

This is why it’s called Retrieval-Augmented Generation. The AI doesn’t just make things up from its training data. It actively pulls in current information and uses that to inform its response.

The retrieval step is where your content either gets selected or doesn’t. The AI is looking for content that clearly matches the query (structural clarity), contains specific information worth citing (factual density), and is confirmed by multiple sources (cross-web consensus). If your content passes those filters, it gets retrieved and potentially cited in the generated answer.

What it means for you

You don’t need to understand the technical details of RAG. But understanding the two-step process (retrieve then generate) clarifies why AEO works the way it does.

Your content needs to be findable (AI crawlers can access it), relevant (it clearly matches the types of questions being asked), extractable (it’s structured so the AI can pull out the specific information it needs), and trustworthy (it’s validated across multiple sources).

RAG is the reason these four things matter. It’s not about gaming an algorithm. It’s about making your content useful to a system that’s actively looking for the best sources to build its answers from.

Related terms: AEO, GEO, Citation, Structural Clarity, Content Freshness

robots.txt: What Is It and Why Does It Matter for AI Search?

In plain English

A simple text file on your website that tells bots (including AI crawlers) which parts of your site they’re allowed to access and which parts are off-limits.

Think of it as the bouncer at your website’s front door. It can let everyone in, block specific visitors, or restrict access to certain areas.

Under the hood

The robots.txt file sits at the root of your website (e.g., yoursite.co.nz/robots.txt) and contains rules written in a standardised format. Each rule specifies a user-agent (which bot the rule applies to) and what that bot is allowed or disallowed from accessing.

For example, a rule might say “all bots can access everything” or “GPTBot is blocked from the entire site” or “PerplexityBot can access blog posts but not product pages.”

AI crawlers check this file before they access your content. If your robots.txt blocks them, they respect that instruction and move on. They won’t crawl your site, which means they can’t cite your content in AI-generated answers.

The problem is that many robots.txt files were written before AI crawlers existed. Default CMS configurations, security plugins, or blanket “block all bots except Googlebot” rules can inadvertently block AI crawlers without anyone realising.

What to do about it

Check your robots.txt file right now. Type your domain followed by /robots.txt into a browser and look at what comes back. If you see rules mentioning GPTBot, ClaudeBot, PerplexityBot, or a blanket disallow for unknown bots, those are blocking AI crawlers.

If you’re not comfortable reading the file yourself, send the URL to your web team and ask: “Are we blocking any AI crawlers?”

This is a two-minute check with potentially significant implications for your AI search visibility.

Related terms: AI Crawlers, Bot Crawl Audit, Log File Analysis

Schema Markup: What Is It and How Does It Help AI?

What it is

A standardised code vocabulary (also called structured data) that you add to your website’s HTML to help search engines and AI systems understand what your content is about.

It’s like adding labels and categories to your content that machines can read.

The mechanics

Schema markup uses a standardised format (from schema.org) to describe things on your page in a way that’s unambiguous to machines. For example, instead of a search engine having to guess that “$49.99” on your page is a price, Schema markup explicitly labels it as the price of a specific product.

Common types include Organisation, LocalBusiness, FAQ, Product, Article, and Review schemas.

For AEO, the most relevant schemas are FAQ (which formats question-and-answer pairs in a way AI can easily extract), LocalBusiness (which provides structured entity information for local queries), and Article (which identifies your content type, author, and publication date).

Why it matters

If your website doesn’t have Schema markup, you’re making AI systems work harder to understand your content.

It’s not a magic bullet. Schema alone won’t get you cited if your content lacks factual density or cross-web consensus. But it removes friction from the process and gives your content a structural advantage over competitors who haven’t implemented it.

Ask your web team whether your key pages have relevant Schema markup, particularly FAQ, LocalBusiness, and Organisation schemas. For multi-location businesses, implementing LocalBusiness schema consistently across all location pages is a significant AEO advantage.

Related terms: Structural Clarity, Modular Content, Google Business Profile

Structural Clarity: What Is It and Why Does It Matter for AI?

The basics

How clearly and logically your content is organised so that AI systems can easily find, extract, and cite specific pieces of information.

It’s about making your content machine-readable without making it unreadable for humans.

Under the hood

AI systems parse content looking for clear signals about what each section covers. Question-based headers (e.g., “How much does X cost?” rather than “Pricing information”) help AI match your content to specific user queries.

Short, focused paragraphs with one clear point each are easier to extract than long blocks of text. Consistent formatting helps AI identify patterns and structure.

The key elements of structural clarity include question-based headers that mirror how users actually search, modular sections that can be extracted independently, clear topic sentences that summarise each paragraph’s main point, and appropriate use of Schema markup to add machine-readable labels to your content.

What to do about it

Audit your most important pages for structural clarity. Can a machine easily identify what each section is about? If someone asked a specific question that your page answers, is the answer clearly findable within a specific section, or is it scattered across multiple paragraphs?

The good news is that content that’s well-structured for AI is also better for human readers. Clear headers, focused sections, and direct answers improve user experience as well as AI extractability.

You’re not choosing between optimising for humans and optimising for machines. You’re doing both at once.

Related terms: Modular Content, Schema Markup, Factual Density, AEO

Third-Party Validation: Why Does It Matter More Than Your Own Content?

In plain English

Evidence from sources you don’t control that supports claims about your brand or expertise.

Reviews, industry mentions, forum discussions, media coverage, and awards all constitute third-party validation. AI systems weight these signals heavily because they’re harder to fake than your own website content.

How it works

AI systems treat third-party validation as a trust signal. Your website saying “we’re the best” carries minimal weight. But if Google reviews, industry publications, Reddit discussions, and professional directories all independently confirm your expertise, the AI’s confidence in recommending you rises significantly.

This is the real-world mechanism behind cross-web consensus.

The platforms that matter most for third-party validation are the ones AI systems already trust and frequently cite: Google reviews, industry-specific publications, major forums and Q&A sites, professional directories, and established news outlets.

The practical takeaway

This is where long-term brand building meets AI search strategy. You can’t shortcut third-party validation. It requires consistently delivering good work, actively generating reviews, building relationships with industry publications, and being genuinely present in the communities where your audience gathers.

Start with the basics: a consistent review generation process, a PR strategy that targets mentions in relevant industry publications, and genuine participation in forums and communities where your expertise adds value.

These aren’t new marketing activities, but they now have a direct connection to AI search visibility that makes them even more strategically important.

Related terms: Cross-Web Consensus, Confidence Threshold, Entity Association, Brand Sentiment, Digital PR

Unlinked Mentions: Why Do They Matter for AI Search?

The short version

References to your brand, products, or people on other websites that don’t include a hyperlink back to your site.

In traditional SEO, these were nearly worthless. For AI search, they’re a significant trust signal.

How it actually works

Traditional search engines built their authority model primarily on links. A mention of your brand on an industry website only counted if it included a clickable link to your domain. No link, no credit, regardless of how positive or prominent the mention was.

AI systems evaluate authority differently. They scan text for entity relationships: how closely and consistently your brand name appears alongside specific topics across the web. Whether there’s a hyperlink attached is largely irrelevant to this calculation.

If an industry journal writes “ACME Corp is one of the leading providers of X” without linking to ACME’s website, traditional SEO gives almost zero credit. An AI system registers that as a meaningful signal about ACME’s association with that topic, especially if similar mentions appear across other independent sources.

The practical takeaway

This changes how you think about PR, content distribution, and partnership activity. The value of getting mentioned isn’t contingent on getting a link anymore.

A podcast interview where the host mentions your brand. A conference write-up that names your speaker. An industry roundup that lists your product. A Reddit thread where someone recommends your service. None of these need to link to your website to build your AI search visibility.

This doesn’t mean links no longer matter. They still do for traditional SEO. But it does mean your PR and content strategy should optimise for mentions and context, not just link acquisition.

Related terms: Entity Association, Cross-Web Consensus, Third-Party Validation, Digital PR

Visibility Decay: Why AI Visibility Fades Without Ongoing Activity

What it is

The gradual decline in how often AI systems cite or recommend your business when new third-party coverage and mentions stop appearing. Unlike traditional search rankings, which can persist for months or years on the strength of existing backlinks and domain authority, AI visibility has a half-life that's directly tied to the recency of your web presence.

How it works

AI systems operate across two layers. The training layer absorbs your accumulated web presence over months and years. The retrieval layer searches the current web in real time when generating answers. Both layers are influenced by recency, but the retrieval layer is especially sensitive to it.

When a user asks a question and the AI searches for sources, recent coverage from the past few months will typically be weighted more heavily than coverage from a year or two ago. Research shows that half of all AI citations come from content published within the previous 11 months. This means your visibility naturally decays as your most recent coverage ages out of that window.

The training layer provides a slower-moving baseline. If your brand has strong entity associations built up over years, those don't disappear overnight. But if the retrieval layer consistently finds nothing recent about you while finding fresh coverage of competitors, the AI will increasingly favour those competitors in its answers.

The decay rate isn't uniform. High-competition categories where multiple businesses are actively building their presence will see faster visibility decay for inactive brands. In quieter niches, your existing presence may sustain you longer. But the direction is the same: without replenishment, visibility fades.

What it means for you

This is the practical reason why a steady cadence of media activity matters more than periodic campaigns. A traditional SEO strategy might allow you to invest heavily in content and links for six months, then coast on those rankings for a year or more. AI visibility doesn't work that way.

The good news is that the activity required to sustain AI visibility doesn't have to be overwhelming. One meaningful piece of contributed expertise per quarter, regular commentary on industry developments, and consistent activity on platforms where your industry conversations happen is enough for most NZ businesses to maintain their presence.

The key is consistency over volume. Four modest but steady touchpoints across the year will sustain visibility better than one large campaign followed by months of silence.

Related terms: Citation Recency, Earned Media Signals, Content Freshness, RAG, Digital PR

Zero-Click Searches: What Are They and What Do They Mean for You?

The short version

Searches where the user gets their answer directly on the search results page without clicking through to any website.

The search engine or AI system provides the information right there, in a featured snippet, knowledge panel, or AI Overview.

How it actually works

Google has steadily added more information directly into search results over the years: weather, calculations, definitions, business hours, and now AI Overviews. When a user searches “what time is it in London” or “how many cups in a litre,” they get the answer immediately. No click needed.

AI Overviews accelerate this trend significantly. When someone asks a more complex question, Google synthesises an answer from multiple sources and presents it at the top of the page.

Approximately 60% of Google searches now end without a click, and as AI integration expands, this could approach 70%.

What it means for you

This doesn’t mean your website doesn’t matter. It means visibility is no longer just about getting clicks.

If your brand is cited in the AI-generated answer at the top of the page, you’re getting visibility and brand recognition even when users don’t click through.

For performance tracking, traditional metrics like organic click-through rate will decline regardless of how well you rank. You need to start tracking brand mentions, AI citation presence, and whether your content is being used as a source in AI-generated answers.

The click might not happen, but the recommendation does, and that recommendation shapes purchasing decisions.

Related terms: AI Overviews, AEO, AI Referral Traffic, Brand Sentiment

Frequently Asked Questions

The Mindset Shift

Three years ago, search was a competition for position. You were trying to be higher on a list than your competitors. Every tactic, every metric, every strategy was ultimately about moving up that list: rank higher, get more clicks, win more traffic.

Now, search is becoming a competition for trust.

The AI systems generating answers aren’t ranking you against competitors on a list. They’re deciding whether to recommend you at all. That’s a fundamentally different dynamic. You’re not fighting for position 3 versus position 7. You’re either in the answer or you’re not. You’re either trusted enough to be cited or you’re quietly skipped.

That shift changes what matters. Position was won primarily on your own website: better content, more links, stronger technical SEO. Trust is earned across the entire web: what your site says, what reviews say, what directories say, what industry sources say, whether all of those tell a consistent story.

You can’t optimise your way to trust with a single channel. You have to earn it everywhere.

The marketing manager who internalises that shift will make better decisions about where to invest their time and budget. They’ll stop asking “how do we rank higher?” as their only question and start also asking “if an AI had to recommend a business in our category, would it confidently recommend us? And if not, why not?”

That’s the question worth sitting with. Because increasingly, it’s the question that determines whether your next customer finds you at all.

Key Takeaways

Consistency is the thread that runs through almost every entry. Entity clarity, Google Business Profile, trust cascade, confidence threshold: they all come back to whether your signals tell the same story across every platform. Inconsistency is the most common reason businesses get quietly skipped by AI systems.

There’s a clear distinction between what you control and what others say about you. Your website, your GBP, your Schema markup: those are yours. But AI systems weight third-party validation, unlinked mentions, reviews, and cross-web consensus more heavily. The entries you control set the foundation. The entries you earn through brand building determine whether you get cited.

Access comes before everything. Bot crawl audits, robots.txt, AI crawlers, log file analysis: four entries all dedicated to the same question. Can AI actually reach your content? None of the trust, structure, or validation work matters if the answer is no.

AI doesn’t rank you. It decides whether to recommend you at all. Several entries (confidence threshold, trust cascade, citation) describe a binary outcome rather than a sliding scale. You’re either in the answer or you’re not. That’s a different dynamic from traditional search, where position 5 still gets some traffic.

The glossary terms work as a system, not in isolation. Factual density without structural clarity means AI can’t extract your data. Cross-web consensus without entity clarity means AI can’t confidently attribute it to you. Schema markup without factual density gives AI a well-labelled page with nothing worth citing. The “Related terms” links at the end of each entry aren’t decorative. They show you which concepts depend on each other.

About This Glossary

This glossary is maintained by Rod Russell at ADMATIC, a search marketing specialist helping New Zealand businesses navigate the shift from traditional search to AI-powered answers. It’s designed for marketing managers who need to understand these concepts without wading through technical documentation.

Missing a term? Got a question about something you’ve encountered? Get in touch and we’ll add it.