LLM (Large Language Model) SEO is the practice of improving how a website, its brand and its content are discovered, understood, retrieved, cited, and recommended by large language model-powered search systems such as ChatGPT, Gemini, and Perplexity.

It extends traditional search engine optimization. The work covers AI search experiences that generate an answer instead of presenting a familiar list of blue links.

There is no settled industry taxonomy or meaning. LLM SEO, generative engine optimization (GEO), answer engine optimization (AEO), LLM optimization and AI SEO (sometimes called AI search optimization) overlap, and different practitioners draw the boundaries in different places. This article uses LLM SEO throughout.

What is LLM SEO?

LLM SEO is the process of improving a brand's visibility in search and discovery experiences powered by large language models. The outcome may be a linked citation, an unlinked brand mention, a product comparison, a recommendation, or a qualified visit to the website. Results shift around. There is rarely a stable "number one position" that holds across every user, prompt and session.

LLM means large language model, the AI model that processes and generates language. An AI search engine may combine that model with a live search layer. That layer finds current documents or passages, and the language model uses the retrieved material to construct a response.

The goal is to move a brand or piece of content through a chain of possible outcomes:

1. Discovery
2. Retrieval
3. Citation or mention
4. Recommendation
5. Qualified visit
6. Conversion

A citation gives the user a source to inspect. A mention can still shape awareness and trust. Recommendations sit further down the journey because the system must understand what the brand offers and who it serves.

The business result is the only measure of success that matters. A hundred AI brand mentions that send no relevant traffic or influence no buying decisions have limited impact on revenue.

How does LLM SEO work?

LLM SEO works by improving the conditions under which an AI search system can access, interpret and select content from a website. The platform's algorithms still decide which sources best support a particular response.

Each stage can fail in many ways. A firewall might block the crawler. The page may be accessible but vague about the entity being discussed. The claim may lack evidence, or another source may explain it more clearly and with more original detail.

How LLMs and AI tools find web pages

LLMs and AI tools find web pages through several routes. A model may use knowledge acquired during training, retrieve live results while answering, or receive documents from a search index through a grounding system.

These routes should not be treated as one universal engine. Google's AI Overviews and AI Mode draw on the Google Search index. ChatGPT Search runs its own targeted searches. Claude runs live web searches when a prompt needs current information. Microsoft Copilot leans on the Bing index, so Bing Webmaster Tools belongs in the setup alongside Search Console. Keep useful content accessible to all of them.

Retrieval-augmented generation (RAG)

Retrieval-augmented generation, usually shortened to RAG, is a method that supplies a language model with retrieved content before it produces an answer. The retrieved content acts as current context and grounds the response in external evidence.

Picture someone asking which accounting platform supports multi-currency invoicing for a UK agency. A retrieval system searches for relevant product pages, documentation and comparisons, selects useful passages and passes them to the model, which writes an answer and may cite those sources.

Google describes RAG in its AI-search guidance as a technique that uses its core Search ranking systems to retrieve relevant, current pages from the Search index. Technical SEO and useful source content still matter here.

Query fan-out

Query fan-out is generating several related queries from one broader request. Those searches help an AI system gather the facts, options and comparisons needed for a complete response.

Ask for the best CRM for a small B2B sales team and the system might investigate CRM features for small teams, pricing for ten users, B2B integrations, setup time, and direct comparisons such as HubSpot versus Pipedrive.

One keyword and one isolated page is a poor model for this kind of retrieval. A stronger site covers the main question and its supporting decisions across a connected structure of pages, with internal links that reflect how people move from one question to the next. Google also warns against creating a separate low-value page for every possible fan-out query. Distinct pages exist because the subjects deserve distinct treatment, not because a tool exported 500 prompt variations.

Ranking eligibility and citation selection

Ranking eligibility and citation selection are different decisions.

Traditional search results may be a good match for a broad query. An AI answer may need one precise statistic, a clean product comparison, a firsthand observation or a passage that supports a particular sentence.

The reverse can happen too. A smaller page that contains a unique, well-supported fact may be useful for one part of a generated response even when it does not rank at the top for the broad head term.

This is a working retrieval model, not a published list of secret citation factors. It explains why conventional rankings, AI mentions and cited URLs should be tracked as separate metrics.

LLM SEO vs SEO, GEO, AEO and AI SEO

The terms overlap, but each marketing label puts the emphasis somewhere different.

Term	Primary focus	Typical outcome
SEO	Visibility in search engines	Organic rankings, impressions, clicks and conversions
LLM SEO	Visibility in LLM-powered search and answer systems	Retrieval, citations, mentions and recommendations
GEO	Visibility inside generative search experiences	Inclusion or citation in generated responses
AEO	Becoming a useful direct answer to a question	Answer surfaces, snippets and AI responses
AI SEO	Broad umbrella for search shaped by AI	Visibility across AI-influenced search experiences

Google acknowledges AEO and GEO as industry terms but states that work on visibility in Google's generative search features remains SEO from its perspective.

LLM SEO vs traditional SEO

Traditional SEO improves visibility in search engines through crawlability, indexing, keywords, content relevance, site architecture, backlinks and user value. LLM SEO keeps those on-page and off-page foundations and adds closer attention to retrieval, brand and entity clarity, source selection, citations, mentions and recommendations. Organic rankings, traffic and clicks remain useful, while AI-search reporting also needs controlled prompt tracking, citation share and mention share. Most marketing dashboards do not show any of this yet.

LLM SEO vs GEO

GEO usually focuses on visibility within generated answers. LLM SEO is frequently used for the same work, though some marketers use it more broadly to cover any discovery or recommendation mediated by a large language model. In practice, both call for accessible pages, original content and credible support. The label on the service matters less than the quality of the work behind it.

LLM SEO vs AEO

AEO grew from the goal of supplying direct answers for featured snippets, voice assistants and answer engines. LLM SEO covers a wider retrieval environment, where an AI system may compare products, combine evidence from several sources or answer a multi-part prompt after running a combination of related searches. Direct answers help, but the surrounding evidence and topic coverage carry equal weight.

What makes content useful for LLM retrieval?

No public checklist can force an AI platform to choose a source, but the following conditions make content eligible, understandable and useful during retrieval.

Make the page crawlable

A crawler must be able to reach the page before its content can be considered. Check server responses, robots.txt rules, noindex directives, canonical tags, internal crawl paths and CDN or firewall settings. Put essential content in readable page text instead of hiding it inside images or scripts.

Platform controls differ, and the platform table later in this article lists the AI crawler each one needs. Crawler access creates an opportunity. It does not secure inclusion, and it is the one thing on this list most sites get wrong.

Answer questions directly

Write headings as plain questions people actually ask and answer each one in the first sentence or two. Then explain the mechanism, provide evidence, add an example and qualify the answer where a real limitation exists.

Avoid turning every paragraph into a rigid 40-word answer block. Google explicitly says tiny content chunks are not required for its generative search features. Natural sections, lists and tables work when they answer a real question and give it enough context.

Remove entity ambiguity

State who or what you are discussing, the relevant attribute and its value in plain words, without forcing the reader to resolve vague references.

For example, "Slack is a workplace communication platform. Salesforce completed its acquisition of Slack in 2021" gives a system two clear facts about one named entity. The copy needs to remove needless ambiguity, nothing more.

Publish original evidence

Commodity summaries give a retrieval system little reason to select one source over another. Add content that came from the work itself: a proprietary study, first-party data, an original screenshot, a benchmark or a documented case study.

A useful LLM SEO article might show a controlled set of prompts tested across three AI platforms, which domains each platform cited and how the results changed over eight weeks. That would give readers insights they cannot get from another generic definition page.

Place evidence beside the claim

Place the evidence beside the claim it supports. Cite the data behind a percentage. Name the researcher, their credentials, the dataset and the date when those details affect interpretation.

A pile of twenty links at the bottom leaves readers guessing about which source supports which sentence, and it makes updates a lot more painful. Point-of-claim sourcing gives the page a cleaner evidence trail that readers and retrieval systems can trust.

Keep brand facts consistent

Conflicting brand descriptions create uncertainty. Audit business names, service descriptions, locations, leadership, dates, prices, product capabilities and policies across all core pages, author profiles, metadata and structured data. Choose a canonical statement for each important fact and update the pages that disagree.

Earn independent corroboration

Independent sources can confirm that a brand exists, does the work it claims and carries authority on a subject. Useful corroboration can come from editorial coverage, industry publications, interviews, independent reviews, trusted directories, LinkedIn, video channels such as YouTube, social media threads and genuine discussions by the people who use the product.

Quality and relevance decide whether those mentions help a reader. Five hundred copied brand descriptions on unrelated sites do not create credible consensus or authority. They are noise, not signals.

Keep time-sensitive facts current

Prices, product features, executive roles, regulations, statistics and availability change. Date these facts where the timing affects their meaning, link to a current source and set a review schedule based on how quickly the information moves: lighter for evergreen definitions, quarterly for a platform comparison, monthly or automated for a pricing table. The visible "last updated" date should reflect a real review, not a cosmetic timestamp change.

A nine-step LLM SEO strategy

Use this nine-step approach to turn an LLM SEO strategy into a repeatable AI search optimization process.

1. Audit AI-search crawl access. Check robots.txt, response codes, canonicals, CDN rules and server logs for the crawlers you intend to allow.
2. Define your central entities and facts. Record the canonical name, description, services, products, people and locations.
3. Map questions and query journeys around the central search intent, then the supporting comparisons and decisions.
4. Group related questions into useful pages. Give each URL a distinct purpose and split topics only when each needs different evidence.
5. Set one clear macro-context for every page. Its title, H1, opening and internal links should agree on the main job.
6. Answer important questions directly beneath the relevant heading, then add explanation, evidence, examples and honest limits.
7. Publish content competitors cannot copy from the SERP. First-party data, original tests, practitioner commentary, screenshots, tools, templates or case studies.
8. Connect the topic with contextual internal links. Build a navigable topic network instead of an orphaned article collection.
9. Earn credible corroboration and measure visibility. Track prompts, sources, mentions, referrals and conversions.

Run the sequence again as the AI platforms change. LLM SEO is ongoing content maintenance, measurement and optimization, not a one-off growth hack.

LLM SEO by platform

Prioritize the platforms your target audience uses and treat their discovery systems separately. Others can wait.

Platform	Discovery route worth understanding	Webmaster consideration
Google AI Overviews and AI Mode	Google Search index, RAG and query fan-out	Meet normal Google indexing requirements and remain eligible for generative AI features
ChatGPT Search	Live search, public web sources and search partners	Allow OAI-SearchBot and OpenAI's published crawler IPs
Perplexity	Web crawling, indexing and on-demand retrieval	Allow PerplexityBot and its documented IP ranges
Gemini	Google grounding and search services where available	Build strong Google Search foundations and current source content
Other LLM assistants	Product-specific models, indexes, search partners or tools	Check the product's current documentation before changing crawler rules

Recheck official documentation before making a sitewide access decision, and verify real requests in server logs instead of assuming a user-agent rule worked.

Does llms.txt matter for LLM SEO?

llms.txt is a proposed text file that gives AI systems a curated, machine-readable guide to a website's important content. It is not a universal requirement for LLM optimization.

Google states that Google Search ignores llms.txt and that publishers do not need special AI text files, AI-specific markup or Markdown versions of their pages.

Perplexity uses it for its own documentation, but that does not make it a cross-platform ranking signal. Create one only when a product you care about documents a real use for it. Spend the core budget on crawlability, original content, evidence and site structure first. Everything else is optional.

How to measure LLM SEO

Measure AI visibility across a fixed prompt set, the sources attached to responses and the traffic and business activity that follows. A single ChatGPT conversation is an anecdote, since answers vary between sessions and users. Use a repeatable prompt panel in a spreadsheet or dedicated tracking tools, record the exact prompt, platform, date, response, cited URLs and brand mentions, and look for movement across several runs and against the competition.

Metric	What it tells you	How to use it
AI mention rate	How often the brand appears across a controlled prompt set	Repeat tests on a defined schedule and segment by platform and topic
Citation rate	How often the domain or a specific page is linked as a source	Record the cited URL and the claim it supports
Share of voice	Brand presence compared with named competitors	Use the same prompts, platforms and testing conditions for everyone
AI referral traffic	Visits arriving from AI search products	Review analytics referral data and platform-specific UTM parameters
AI conversion rate	Whether AI-referred visitors become leads or customers	Compare landing pages, intent and conversion quality with other channels
Crawl activity	Whether relevant AI-search crawlers access the site	Inspect server logs by verified user agent and IP range
Google generative AI impressions	Links to the site shown in Google's supported generative features	Use Search Console's Generative AI performance report by page, date, country and device

Google's Generative AI performance report covers AI Overviews and AI Mode and was rolled out worldwide by Aug 31, 2026. Search Console may withhold or aggregate some data, so read the report as one part of the measurement set.

OpenAI says ChatGPT adds utm_source=chatgpt.com to referral URLs. That gives analytics teams a clean way to isolate inbound traffic from ChatGPT Search.

Common LLM SEO mistakes

The biggest mistakes come from strategies that treat a changing retrieval environment as a bag of shortcuts.

• Separating LLM SEO from the technical and editorial foundations of SEO.
• Publishing hundreds of thin AI-generated pages for slight prompt variations.
• Blocking a search crawler at the CDN while allowing it in robots.txt.
• Treating llms.txt as a universal ranking hack.
• Adding schema markup because of a vague claim that "AI needs schema."
• Repeating familiar claims without first-party evidence or original analysis.
• Measuring one response and calling it an AI ranking.
• Manufacturing reviews, directory entries or brand mentions to fake consensus.

[fdb_faq title="Frequently Asked Questions"]

[fdb_faq_item question="Does LLM SEO replace traditional SEO?" open="true"]

No. LLM SEO builds on technical SEO, crawlability, site architecture, content quality, user experience, authority and measurement. AI search adds new retrieval and answer formats, but all of it still needs accessible sources.

[/fdb_faq_item]

[fdb_faq_item question="Can LLM SEO guarantee placement in ChatGPT Search?"]

No. OpenAI says placement in ChatGPT Search is not guaranteed. You can improve eligibility by allowing OAI-SearchBot, making content accessible and publishing material worth citing, but the system selects sources for each response.

[/fdb_faq_item]

[fdb_faq_item question="Does structured data improve LLM SEO?"]

Structured data can clarify page information and make a page eligible for certain traditional rich results when the markup follows platform rules. Google says it is not required for its generative AI search features, and there is no special generative AI schema type. Do not add unsupported markup to chase an assumed AI ranking benefit.

[/fdb_faq_item]

[fdb_faq_item question="Can I block GPTBot but allow OAI-SearchBot?"]

Yes. OpenAI separates GPTBot, which publishers can disallow to exclude pages from potential model training, from OAI-SearchBot, which supports discovery and inclusion in ChatGPT Search. Configure and test both user agents separately.

[/fdb_faq_item]

[fdb_faq_item question="Is AI-generated content bad for LLM SEO?"]

AI-generated content is not automatically bad for LLM SEO. Low-value, inaccurate or mass-produced content creates the problem, and Google warns that scaled pages with little user value may violate spam policies. Use AI for research and drafting as part of a real editorial process. Keep humans responsible for expert judgment, verify every factual claim and contribute evidence that did not come from summarizing the existing results page.

[/fdb_faq_item]

[fdb_faq_item question="Can small businesses compete in LLM SEO?"]

Yes. Public eligibility is not limited to the largest brands or domains. A smaller site can supply a specific, original and well-supported fact that helps answer a question. Small businesses should focus on narrow expertise, original evidence and clear topical connections instead of trying to imitate the breadth of a large publisher.

[/fdb_faq_item]

[/fdb_faq]

LLM SEO starts with a simple question: when an AI system needs evidence about your brand or market, what content does your website contribute that deserves to be retrieved? If the honest answer is very little, that gap is the work. Make that contribution accessible, specific and easy to verify, then measure whether it earns citations, qualified visits and customers.

An AI visibility audit tells you whether ChatGPT, Google AI Mode, Perplexity, Gemini, and other answer systems mention, cite, or recommend your business when a customer asks a question you should win.

Give me an afternoon, a spreadsheet, and a private browser window. By the end, you will know how often AI names you, which competitors it names instead, the gaps you have and which sources it trusts, and whether your result survives a second or third run.

When we launched the free AI Visibility Scorecard, 1,500 businesses ran it in the first week. Eighty-nine percent were invisible to AI. Most had no idea.

You cannot fix a visibility problem you have never measured. Start here.

What is an AI visibility audit?

An AI visibility audit is a practical assessment of how AI search systems represent your business. Use the first pass as an evaluation of your baseline. It answers one question: when a potential customer asks AI platforms who to buy from, hire, or use, does your name come up?

It is the AI-search equivalent of checking Google rankings, with two differences. There is no fixed position. An AI answer is a paragraph, and you are either in it or you are not. The answer can also change when you run the same prompt again, so you measure a rate of appearance rather than a permanent spot.

The audit has four jobs:

Choose the prompts your customers use.

Run those prompts across the models that matter to your market.

Log the answer, the brands named, and the sources cited.

Diagnose the gap and rerun the same test later.

Keep three outcomes separate. A mention, a citation, and a recommendation are different signals.

All help you understand the sentiment, authority, and presence your brand has when AI crawlers assess your content and online findings.

Mentions, citations, and recommendations measure different things

A mention is your name appearing anywhere in the answer. “Other real estate agents include Brand A and Brand B” puts you in the paragraph. That is all it proves.

A citation is a link to a page attached to the answer as a source. The model found a page it could use and chose to show that page to the reader. The citation may point to your website, your Zillow profile, Realtor.com, a local publication, or another page that describes you.

A recommendation is the commercial outcome. The model names you as a pick, gives a reason, and connects you to the customer’s situation. Weight recommendations more heavily in your notes because a passing mention and a qualified referral are not equivalent.

A source can influence an answer without getting a visible link

A source is any page the system retrieves or reads while building an answer. A citation is the subset of those pages it shows as a link.

Ahrefs analyzed 1.4 million ChatGPT 5.2 prompts from February 2025 and reported that ChatGPT cited about half of the URLs in its retrieval pipeline. Its study also found that 88.46% of the URLs in the search ref_type were cited, while dedicated Reddit, YouTube, and academia channels behaved differently. The data is observational, not a published OpenAI ranking formula, but the operational lesson is clear: being retrieved and being cited are separate events. Read the Ahrefs study.

If ChatGPT recommends you and links to your G2 profile, G2 received the visible citation while you received the recommendation. Log both. The pair tells you where the system found evidence about your business and which page it trusted enough to expose.

How to find your current AI visibility score

Run five prompts first. Ten minutes are enough to tell you whether the full audit will measure a small gap or a crater.

Open ChatGPT in a private browser window, log out, and ask questions that match your market and the topics you want to be in the responses for.

For a Newton real estate agent, that might be:

“What is Newton Harbor Realty?”

“Who is the best listing agent in Newton, Massachusetts for a seller with a $2 million home?”

“I am buying my first home in Newton with a $500,000 budget. Which agent should I speak to?”

“Is Newton Harbor Realty a good choice for selling a home?”

“Newton Harbor Realty reviews.”

To effectively run an AI visibility audit, you should start by analyzing the prompts users commonly enter, as these reveal the questions and concerns your AI needs to address most accurately. Replace the example business and market with your own. Repeat the five prompts in Google AI Mode and Perplexity, then save the answers rather than trusting your memory.

A wrong description, a discontinued service, or “I do not have information about that” points to an entity problem. AI does not have a stable picture of who you are.

A correct branded answer followed by competitors on category and problem prompts points to a corroboration problem. AI knows you exist but has less evidence for recommending you.

Showing up across all fifteen results is useful, but it is still a baseline. Run the full audit before you call the result durable.

Decide What Goes Into the Audit

An audit is only as useful as its prompts. Define the customer question, the models, the location, and the exact prompt set before you open the spreadsheet.

Choose a customer question narrow enough to score

“Real estate agent” is not a target. “Luxury real estate agent for homes above $10 million in Austin” is a target. The tighter definition gives the model a clearer connection between your business, your customer, your service, and your market.

Write down three to five target questions. Each should combine what you do, who you do it for, and where the service applies. Use the words buyers and sellers use, not the language on your internal strategy deck.

If your team calls the service a “revenue intelligence platform” while buyers ask for “sales forecasting software,” the customer language belongs in the prompt list. Your existing keyword research is a useful starting point. If you have not done it, start with the difference between AI SEO and traditional SEO before you build a tracking system.

Test the models your customers can reach

Start with four surfaces. Add Claude, Copilot, and Grok after the process is stable, or move one of them earlier when your customers use that product heavily.

Model or surface	Why include it	First-pass setup
ChatGPT	Tests a conversational AI-search experience with web sources.	Private window, logged out.
Google AI Mode and AI Overviews	Tests the generative layer attached to Google Search.	Record market and device context.
Perplexity	Shows source links inline, which makes source mapping easier.	Use the same location and prompt wording.
Gemini	Adds a Google ecosystem surface to the comparison.	Keep the account state consistent.

Google says AI Overviews and AI Mode may use query fan-out, where several related searches across subtopics and data sources help form one response. Google also says the same foundational SEO practices apply: pages need to be indexed and eligible for a snippet, and there are no special AI-specific technical requirements. Read Google’s current guidance for AI features.

Build the prompt list around four customer situations

Your prompt list is the audit. Build 20 to 30 prompts, with more category and problem prompts than branded prompts. Those are the prompts closest to a new customer choosing who to call.

Prompt category	Example	What it tests
Branded	“What is Newton Harbor Realty?”	Entity recognition and factual accuracy.
Category	“Who are the best listing agents in Newton for 2026?”	Recommendation for the service and location.
Comparison	“Compass versus Keller Williams agents in Newton.”	Whether you enter the consideration set.
Problem	“I need to sell my Newton home within 60 days. Who should I call?”	Match between your footprint and a real situation.

Add branded prompts such as pricing, reviews, service area, and years in business. Add category prompts for buyer’s agents, listing agents, probate sales, relocation, and luxury homes. Add comparison prompts that name the brokerages your local clients already consider. Add problem prompts with a time limit, neighborhood, property type, budget, or other constraint.

Write the prompts the way a real person types them. “Best listing agent Newton MA” and “I am selling a three-bedroom house in Newton and need someone who knows the local market” are both valid. One is compressed. The other is conversational. Your audit should see both.

Choose the Right Audit Tool for the Stage You Are In

You do not need to buy anything for the first audit. A spreadsheet and the free versions of the models are enough for an initial analysis.

Manual testing teaches you what the tool hides

Type each prompt, read the whole answer, and log the result. Twenty-five prompts across four models, run three times each, produces 300 answers and roughly four hours of work.

That is tedious. It is also the best way to understand your market because you see every recommendation, wrong description, competitor, and source with your own eyes. Once the baseline is clear, automation can improve efficiency without replacing judgment.

Free checkers give you a snapshot. Ahrefs has a free AI visibility checker for a small set of prompts, and the FlyDragon AI Visibility Scorecard is another quick starting point. Use either to decide whether a full audit is worth your afternoon.

Tracking tools automate the prompt list and support ongoing monitoring, so you can compare performance without changing the baseline. Ahrefs Brand Radar, Profound, Peec AI, and Semrush’s AI toolkit all sit in this category. Before you pay, judge each tool’s effectiveness by whether it preserves the prompts, model context, and raw answers. The source draft places typical costs between $100 and $500 or more per month; check current pricing before you commit.

Ask whether the tool uses API calls or the customer interface

API calls send a prompt to a model endpoint and save the response. They are fast and easy to scale, but the endpoint may use a different model version, omit web search, or lack the account and interface context your customers experience.

Browser emulation drives the consumer interface, enters the prompt, and captures the answer. It is slower and can break when the interface changes, but it is closer to a customer opening ChatGPT or Perplexity on a phone.

Ask the vendor which method it uses, whether web search is enabled, which model version it runs, and how it handles location. If the vendor cannot answer, treat the data as a different measurement surface. When a manual result and a tool result disagree, do not average them together.

Run the Audit With a Controlled Setup

Block off an afternoon. Put the prompt list and the spreadsheet side by side. Use the same setup for every run.

Log out of every model for the baseline. A logged-in session can carry history, memory, or account context that a new customer will not have.

Record the location. For local work, run from the market you serve or use a consistent test location. For national work, choose one location and keep it constant.

Run every prompt three times. One answer is an observation. Three answers give you a rate.

Check crawler access as a separate technical task, including a basic robots.txt compliance check. Google says eligibility for AI Overviews and AI Mode requires a page to be indexed and eligible for a normal Search snippet. OpenAI says publishers should avoid blocking OAI-SearchBot if they want public pages included in ChatGPT summaries and snippets. Perplexity says PerplexityBot follows robots.txt and will not index full or partial text where the site disallows it. Those controls establish access; they do not guarantee a recommendation. Read OpenAI’s publisher guidance and Perplexity’s robots.txt guidance.

Use a 0-to-3 score for each answer

Score	Meaning	Example
0	Absent	Your business does not appear in the answer.
1	Mentioned	Your name appears without an endorsement.
2	Cited	A page about you is linked as a source.
3	Recommended	You are named as a choice with a reason.

Scores overlap by design. If you are recommended and a page about you is linked, record the highest outcome, 3, and save the citation in the source column. If you are mentioned and a page about you is linked without a recommendation, record 2.

Your spreadsheet needs eight columns: Date, Model, Prompt, Category, Run number, Score, Brands named, and Sources cited. Those columns give you the core metrics for comparing models and prompt categories over time.

A sample row dated 2026-09-02 might read: ChatGPT; “Best real estate agent in Newton MA”; Category; Run 1; Score 0; J. Doe from Compass, M. Lee from Keller Williams, K. Park from Redfin; Zillow.com, Realtor.com, Reddit.com.

Save the last two columns every time, including a zero. “Brands named” becomes your competitor list. “Sources cited” shows you where AI learns about the category.

Read the Results as Rates, Not Anecdotes

Turn the rows into an overall rate, a rate per model, and a rate per prompt category. Keep a benchmark, competitor list, and source list beside those metrics so the analysis stays useful month to month. This creates a clean basis for monthly benchmarking. Your visibility rate is total points divided by maximum possible points.

For 25 prompts, three runs, and a maximum score of 3, the maximum is 225 points per model. A score of 41 is 18%. Calculate the rate per model and per category because they answer different questions.

Slice	Score	Maximum	Rate
Overall	112	900	12%
ChatGPT	41	225	18%
Google AI Mode	38	225	17%
Perplexity	24	225	11%
Gemini	9	225	4%
Branded prompts	67	135	50%
Category prompts	22	315	7%
Comparison prompts	8	135	6%
Problem prompts	15	315	5%

The table is an illustrative worksheet, not a benchmark for every business. It tells a useful story: the business is recognized when named and rarely recommended for an unbranded customer need. That points to corroboration and category coverage. A different business with 8% on branded prompts and 8% across the other categories has an identity problem first.

Why the same prompt produces different answers

Variation is normal. Models sample from probabilities, retrieval algorithms can change the source set, and the open web changes underneath the test.

Query fan-out adds another source of movement. Google documents that one complex prompt can become several related searches. OpenAI also describes targeted search queries and follow-up searches in its ChatGPT Search material. A different sub-question can change the pages retrieved, the competitors found, and the sources shown.

Location, account state, model versions, new competitor pages, updated reviews, and a new Reddit thread can move the result. A September answer will not be a perfect copy of an August answer.

Three runs do not remove every source of variance. They give you a first estimate. The paper Don’t Measure Once: Measuring Visibility in AI Search argues that one-off observations are unreliable because AI answers vary across runs, prompts, and time. A snapshot is a lead. The pattern is the measurement.

If a prompt scores 3, 0, and 3, your recommendation rate for that prompt is 67%. Record the pattern rather than choosing the most flattering answer.

Use the Competitor and Source Lists to Find the Gap

Sort “Brands named” by frequency. The top five names are your competitors in AI search, and they may not match your sales team’s battlecards.

Run branded prompts for each competitor. Note what AI says about the business and which pages it cites. You are looking for the evidence attached to the name.

A specific, verifiable claim, such as 52 transactions in Newton in 2025 or a 4.9 Zillow rating from 210 reviews.

A third-party page, local news item, brokerage announcement, or genuine past-client recommendation that names the business beside the category.

Agreement across the website, Zillow, Realtor.com, Google Business Profile, LinkedIn, the brokerage page, and the state license lookup.

A page that states plainly what the business is and who it serves.

Then sort “Sources cited” by frequency. Check whether you appear on those pages, whether the information is accurate, and whether the profile describes the service you sell today. A source map gives you useful insights into what to fix, pitch, update, or create.

Read how real estate agents protect their brand in LLMs and AI search when the source list exposes a wrong brokerage, city, service, or review story.

Diagnose Why You Are Not Showing Up

Low scores usually point to one of five causes. Rank them before you start making changes.

Your entity is unstable. Branded prompts fail or return conflicting answers because your name, category, service area, and key facts disagree across profiles, or another business has a stronger claim to the name.

Your own site is the only page saying you are good. Branded prompts pass, category prompts fail, and third-party sources provide little corroboration.

Your content is attached to the wrong use case. You appear for an old service, an adjacent category, or a broad term that does not match the customer’s wording.

The important information is difficult to access. JavaScript-only rendering, blocked crawlers, images containing core facts, or PDFs holding the only service details can keep useful evidence out of the retrieval path.

You are described with adjectives instead of definitions. “Passionate about helping families find their dream home” gives a model little it can safely reuse. “Jane Smith is a listing agent in Newton, Massachusetts who has sold 200 homes since 2015” gives it an entity, service, place, and verifiable number.

Most businesses have two or three causes at once. Fix the one that blocks the rest. A perfect schema implementation will not rescue a profile that names the wrong city.

Improve AI Visibility From the Audit

Work through the cheapest fixes first. Treat each enhancement as a response to a failed prompt, a missing source, or a factual conflict.

Make every public profile agree

Use the same name format, one-line description, category, service area, pricing language, phone number, and logo on Zillow, Realtor.com, Google Business Profile, LinkedIn, Homes.com, Yelp, your brokerage page, and the state license lookup. Update the source pages that appeared most often in your audit before creating another profile.

Open with a definition

Your homepage, About page, and profile bios should begin with a sentence a model can lift without rewriting: “Business Name is a category for customer type that does differentiator, with verifiable number since year.” Replace the generic terms with your facts. Then tell the story.

Get named by pages you do not control

Use the “Sources cited” list to choose the right places: a local news story, a brokerage release, a market report, a podcast transcript, a neighborhood guide, or a genuine client recommendation. You need a page that names you beside your category and customer, not another self-description on your own website.

Publish for the prompts you lose

Every problem prompt with a zero is a page opportunity. “Relocating to Newton with children” is one page. “How to sell a Newton condo quickly” is another. Match the title to the customer question, open with the answer, and use real streets, neighborhoods, sale prices, property types, and numbers where they are relevant and verifiable.

Make the site readable

Check that important text appears in the HTML, that your pages are accessible to the relevant crawlers, and that your core facts are not trapped inside images or PDFs. Use structured data when it accurately describes the visible page; treat it as an optimization aid, not a substitute for clear copy. Google says there is no special AI schema or machine-readable file required for AI Overviews or AI Mode, so do not buy a markup package that promises a secret shortcut.

If you plan to hire an agency for this work, ask how it measures visibility, what it knows versus what it is testing, and whether it can show the prompt set and raw answers. Use these 12 questions before you sign an AI SEO agency.

Monitor the Trend Without Ruining the Baseline

Keep the original prompt list, the same model set, the same location, the same account state, and the same three-run method. Add a Month column and rerun the full audit on a schedule.

Monthly is enough for a manual audit. Weekly tracking makes sense when a tool is already running the prompts for you. The changes that matter—a corrected profile, a third-party mention, or a newly indexed page—usually take weeks to appear in answers, while a daily manual check mostly records noise.

Watch the Brands named list for a new competitor appearing three months in a row. Watch the Sources cited list for a review site, local publisher, Reddit thread, or directory that keeps entering the answer. Those lists tell you where the market is moving.

Expand the prompt set as the business grows, but keep the original set intact so the trend line survives. Treat that original set as your benchmark for later optimization. If a tracker alerts you to a drop greater than ten points week over week, inspect the raw answers before changing the strategy.

How often should you rerun the audit?

Run it monthly by hand and weekly with a tracking tool. The measurement research in the FlyDragon source library repeatedly treats AI visibility as a distribution across prompts, runs, platforms, and time rather than a single score. Its practical warning is simple: a larger number of identical repeats cannot compensate for a prompt list that does not represent the customer’s real language.

[fdb_faq title="Frequently Asked Questions"]

[fdb_faq_item question="How often should I run an AI visibility audit?" open="true"]

Run it monthly when the process is manual and weekly when a tracking tool runs it for you. Keep the original prompts and setup unchanged so the comparison remains valid.

[/fdb_faq_item]

[fdb_faq_item question="Why do I get different answers when I run the same prompt twice?"]

AI systems can sample different wording, issue different retrieval searches, receive different source sets, and operate on changing model versions. Run each prompt three times and score the pattern. Two recommendations out of three is a 67% rate; one run is an anecdote.

[/fdb_faq_item]

[fdb_faq_item question="Do I need to be logged out when I test?"]

Use a logged-out private window for the baseline. A logged-in session can carry history, memory, or account settings that distort what a new customer sees. Run a separate logged-in test only when you want to study a returning customer’s experience.

[/fdb_faq_item]

[fdb_faq_item question="Which AI models should I test first?"]

Start with ChatGPT, Google AI Mode or AI Overviews, Perplexity, and Gemini. Add Claude, Copilot, and Grok when the core process is stable or when your customers use those surfaces often.

[/fdb_faq_item]

[fdb_faq_item question="Does an AI visibility audit replace SEO?"]

No. Google’s guidance says AI features use the same foundational SEO practices, and Ahrefs found that the general search channel made up 88% of ChatGPT’s cited URLs in its study. SEO helps pages enter the retrieval pool. The audit shows what happens after your pages and profiles are available to the system.

[/fdb_faq_item]

[fdb_faq_item question="What is the difference between being mentioned and being cited?"]

A mention is your name in the answer. A citation is a source link attached to the answer. A recommendation names you as the choice and gives a reason. Log all three because each one points to a different problem or opportunity.

[/fdb_faq_item]

[fdb_faq_item question="Why am I invisible even though I rank on Google?"]

Ranking can make a page eligible for retrieval, but it does not guarantee that the page will be selected, used, cited, or turned into a recommendation. Check whether your name, category, service area, and proof agree across the pages the audit finds.

[/fdb_faq_item]

[fdb_faq_item question="How long does it take to see improvement after fixing issues?"]

Plan for four to twelve weeks. In our client work, the average time to a first AI mention after profile cleanup and third-party corroboration is about six weeks. Re-audit monthly and expect branded prompts to move before category and problem prompts.

[/fdb_faq_item]

[/fdb_faq]

Run the first fifteen prompts this afternoon. Save the answers. The second run is when the audit becomes useful.

This page is the best guide on AI visibility audits. Relevant terms to AI visibility are: agencies, search engines, signal, comparison queries, txt file, brand sentiment, opportunities, platform, conversations, benchmarks, structure, backlinks, impact, wins, volume, schema markup, llm

AI SEO is the practice of improving how a website, brand, or information source is discovered, understood, retrieved, and represented in AI-powered search. 

It builds on traditional SEO, with one big difference in what you're optimizing for — these search experiences generate answers, summarize information, cite sources and recommend entities. 

They don't just hand you a ranked list of webpages anymore.

Now, the term itself is a bit of a mess. Nobody agrees on one definition yet.

In this guide, AI SEO means one thing: optimizing for visibility inside AI-powered search and answer experiences. When we're talking about using AI to do SEO tasks (keyword research, content generation, analysis), we'll call that AI-assisted SEO.

Using ChatGPT to write an article and making your company relevant, accessible, and well-supported enough to show up when someone asks an AI about your market are two completely different jobs. 

People conflate them constantly.

If you're evaluating providers rather than learning the definition, our guide to choosing the best AI SEO agency for real estate agents covers that decision separately, alongside the questions to ask an AI SEO agency before you sign.

What Does AI SEO Optimize For?

AI SEO optimizes for a wider set of visibility outcomes than a conventional ranking.

Traditional SEO asks one question: 

Can this page be discovered, indexed, understood, and ranked for the relevant search? 

AI SEO keeps that question and adds a second one:

Can the information, source, or entity be retrieved and usefully represented when an AI system generates an answer?

Depending on the query and the platform, "represented" could mean a citation to a webpage, a supporting link, a summarized fact, a company or product mention, a local business, or a straight recommendation.

None of this means webpages have stopped mattering. Google says its generative search experiences remain rooted in its core Search ranking and quality systems, with relevant webpages pulled from the Search index before anything gets used to build a response. 

OpenAI says much the same about ChatGPT Search — public websites can appear in results, and publishers who want their content discoverable, surfaced and cited should allow its OAI-SearchBot crawler.

Webpages are still central. What's changed is the shape of the journey. You used to optimize for one path:

Query, ranked page and click

Now you also need to understand this one:

Question, retrieval, supporting information, generated answer, source, entity or recommendation.

How Does AI SEO Work?

AI SEO works by improving the information available at several points in the search and retrieval process.

Start with discoverability. Your content has to be reachable in the environment you're targeting. For Google, a page must be indexed and eligible to appear with a snippet in normal Search before it can qualify as a supporting link in AI Mode or AI Overviews. And Google says there are no extra technical requirements for those AI features beyond that.

Then the system has to work out what the user is asking for, which can go well beyond matching the literal words in the question. Google publicly documents a technique called query fan-out, where AI Mode and AI Overviews fire off multiple related searches across subtopics and data sources to build a response to a complicated question. One question in, many searches out.

From there, relevant pages, passages, data, or entities get retrieved. The system weighs what it found and generates a response. Supporting pages or businesses may get surfaced alongside that response, depending on the feature and the query.

So the practical AI SEO job is making the right information discoverable, relevant, unambiguous, useful, well-supported, and connected to the right entity or source. Which is why this is bigger than dropping a new keyword into an article.

Compare these two searches. First: "best real estate agent." Second: "Who is the best real estate agent to sell my $500,000 condo, by the beach, with at least 50 reviews?”

The second question carries a specific requirement and comparative intent all in one go. A source that can only answer the broad category question may not be the best source for the constrained version.

What information can AI-powered search use?

There's no universal source list shared by every platform. Google retrieves through Google Search. Other AI search products run on different search providers, crawlers, indexes, and models. The source mix can even change based on the question itself.

Which is why "rank in ChatGPT" is an oversimplification. AI visibility is a search ecosystem, not a single position (and you should think about it that way from the start).

Why Is AI SEO Important?

AI SEO matters because AI-powered search lets people express far more of their real problem in a single interaction.

Google describes AI Mode as “especially useful for nuanced questions, complex comparisons and queries that used to take multiple searches”. Its query fan-out system can pull information across multiple related subtopics before building the response. That changes how much context lives inside one search journey.

Someone choosing an agent to work with can specify team size, location and budget.

Traditional keyword search isn't going anywhere. But more expressive queries mean more attributes deciding whether your business is relevant — which makes information architecture more important, not less. 

You can't assume that ranking for the broadest market term means you've answered every more specific version of the user's need. In my experience, most companies haven't come close.

AI search can also change the outcome you're competing for. Sometimes the goal is still a click. Sometimes the brand needs to be included in the answer. Sometimes your research needs to become the cited evidence. And sometimes the business itself is the entity being compared or recommended.

Is AI SEO Different From Traditional SEO?

AI SEO is an extension of SEO, not a replacement for it. Technical accessibility, relevance, useful content, information architecture, and authority all still matter. AI-powered search just adds new visibility outcomes on top: generated answers, citations, entity selection.

Google is unusually direct on this point. Its current guidance says SEO best practices remain relevant to AI Mode and AI Overviews, and its newer generative AI optimization guidance says those features are rooted in Google's existing Search ranking and quality systems.

[fdb_table title="Traditional SEO vs AI SEO"]

Dimension	Traditional SEO	AI SEO
Primary visibility outcome	A webpage ranks in search results	A page, source, fact, entity or brand becomes part of an AI-powered search experience
Typical query	Often concise or keyword-led	Can include longer conversational questions and multiple constraints
Main retrieval object	Web documents and search features	Web documents still matter, but information may be synthesized into a generated response
User decision	Often made after visiting search results	May begin inside the generated answer before a website visit
Measurement	Rankings, impressions, clicks, organic conversions	Traditional metrics plus AI citations, mentions, recommendation share, AI referrals and downstream conversions
Technical foundation	Crawlability and indexability	Still important; requirements depend on the targeted platform
Content objective	Become the strongest page for a search need	Become a strong page/source and make its information easy to retrieve within the relevant context

[/fdb_table]

The distinction is useful for planning, but don't split the two into separate departments. If a search system can't access a page, the quality of its copy is irrelevant to that retrieval. If a page is accessible but full of commodity information, accessibility alone won't make it a source worth citing. And if a brand is mentioned everywhere but the underlying facts are inconsistent, more visibility just spreads the inconsistency further.

What Is the Difference Between AI SEO, AEO, GEO and LLMO?

These terms overlap heavily, and the industry hasn't standardized any of them. Treat them as overlapping ways of describing different parts of modern search optimization rather than four separate technical disciplines — because that's what they are.

Even Google now acknowledges that AEO and GEO are common industry terms, while maintaining that from its perspective, optimizing for its generative AI search experiences is still SEO.

[fdb_table title="AI SEO, AEO, GEO and LLMO"]

Term	Common meaning	Primary emphasis
SEO	Search Engine Optimization	Organic search visibility
AI SEO	SEO adapted to AI-powered search and answer experiences	Broad umbrella covering search + AI visibility
AEO	Answer Engine Optimization	Becoming a useful direct answer
GEO	Generative Engine Optimization	Visibility within generated AI responses
LLMO	Large Language Model Optimization	Being understood, retrieved or represented by LLM-based systems
AI-assisted SEO	Using AI to carry out SEO work	Production and workflow rather than search visibility

[/fdb_table]

The labels are handy for communicating different objectives. But swapping the acronym doesn't create a new set of search fundamentals. A company with weak information, poor accessibility, and no meaningful evidence doesn't fix any of that by calling the work GEO instead of SEO. I've watched plenty of agencies try.

What Are the Main Components of AI SEO?

A practical AI SEO strategy breaks down into seven connected components: technical discoverability, entity clarity, query coverage, information quality, original evidence, external corroboration, and measurement.

To be clear: this is a planning framework, not a claim that every AI platform runs seven ranking factors or weights them the same way.

1. Technical Discoverability

Technical discoverability decides whether a search system can reach the information you want it to retrieve.

For Google's generative search, normal Search eligibility is the foundation: the page must be crawlable, indexed and eligible to appear in Search with a snippet. Google specifically recommends making sure crawling is allowed, important content is available as text, internal links make pages findable, and relevant Business Profile or Merchant Center information stays current. 

OpenAI separately recommends allowing OAI-SearchBot when you want your content included in ChatGPT Search summaries and snippets.

Most people misunderstand that crawler controls are platform-specific. There's no universal switch labeled "allow AI." So an AI SEO technical audit starts by identifying which systems matter to your business and checking whether the information those systems need is publicly accessible in a format they can use.

2. Entity Clarity

Entity clarity means making it easy to understand exactly who or what a page is describing, and how that entity relates to other entities. An entity can be a company, person, product, location, or organization. Anything identifiable.

Picture a brokerage that recently changed its name, address, and phone number. Its homepage uses the new name. 

Old review sites use the previous brand. LinkedIn describes an earlier product category. Third-party articles list pricing that no longer exists. Those look like branding inconveniences, but from a retrieval perspective they're conflicting assertions about the same entity — and conflicting assertions are poison for AI visibility.

The goal here is not to repeat the entity name over and over. You're maintaining consistent, explicit facts about what the entity is, everywhere those facts appear.

3. Query and Topical Coverage

Query coverage means building enough relevant information to satisfy the important questions and subtopics around a search need without spinning up a separate page for every possible wording.

(To clarify, this means if you ever receive advice to publish hundreds of FAQ pages to your site, it’s not helping AI search.)

A broad query usually represents a network of follow-up questions. Someone researching "homes in Florida" will probably go on to ask about house values, transport, schools, and nearby alternatives.

So AI SEO benefits from designing an information architecture around entities, attributes, questions and relationships — rather than churning out isolated articles from an unordered keyword list. 

But broader coverage doesn't mean unlimited URLs either. 

Google explicitly says publishers don't need to break content into tiny chunks for generative AI or rewrite pages to capture every variation of a long-tail query. Its systems can interpret synonyms and the general meaning behind searches.

The better approach: work out which context deserves a canonical page, answer it properly there, and connect narrower or adjacent contexts where separate pages genuinely make sense.

4. Information Quality and Answerability

Information quality decides whether the retrieved content resolves the user's need accurately, clearly, and efficiently.

Don't write AI SEO content as disconnected "LLM chunks", Google specifically says chunking content purely for generative AI is unnecessary. Let the format follow the question instead. 

• A definition deserves a definitive paragraph. 
• A comparison may need a table, 
• A process usually reads clearer as ordered steps, 
• A numerical relationship may need a chart, 
• A physical process may need a diagram or video, 
• And a complex claim may need supporting research behind it.

Page-level architecture matters here too. The first paragraphs should establish what the page answers. Each section after that should deepen or logically extend that context, rather than hopping between loosely related keywords because a spreadsheet said so.

Clarity serves humans first. It also strips out ambiguity for any system trying to retrieve or summarize the information.

5. Original Information and Evidence

Original information gives a source something worth retrieving that can't be reproduced by summarizing the existing SERP.

That can be first-party research, experiments, original datasets, expert observations, proprietary methodologies, case studies, calculators, firsthand testing or genuinely new analysis. 

Google's current generative AI guidance specifically recommends creating non-commodity content, and says unique, compelling, useful information is likely to influence long-term presence in generative search more than generic optimization tricks. 

Its broader content guidance asks the same question in a different way: does this page contain original information, reporting, research or analysis beyond what's obvious elsewhere?

This is the difference between covering a topic and being a source on a topic. Anyone can summarize ten existing articles in an afternoon. The original experiment those ten articles end up citing is much harder to replace and that's exactly what you want to be.

6. External Corroboration and Information Consistency

External corroboration is evidence outside your own website that supports or qualifies claims about the entity being searched.

How much it matters varies by topic, query and platform, and no AI search engine publishes a universal weighting for "brand mentions" or third-party sources. 

But the underlying information problem is simple. 

A business claiming to be the market leader is making a first-party assertion. An independent industry dataset showing it has the largest market share is different evidence entirely. A doctor's own website listing a qualification is one source; the professional register confirming that qualification is another. A manufacturer's spec sheet is first-party; independent testing gives a second perspective.

So AI SEO has to consider both what a source says about itself and whether the important claims hold up across the wider information environment. Google explicitly cautions against chasing inauthentic mentions as an AI-search hack. You want corroboration, not manufactured claims.

7. Measurement and Refreshment

AI SEO has to be measured over time, because visibility shifts by query, platform, source and date. Traditional SEO metrics stay useful. Rankings, organic impressions, traffic and conversions tell you whether the underlying search visibility is improving. 

AI-specific measurement adds its own questions: 

• Does the brand appear in relevant answers? 
• Is it cited? 
• Which URL gets used? 
• What claim does the cited page support? 
• Does the result hold when the prompt changes? 
• Is the brand present across multiple platforms? 
• Does any of it produce referral traffic, branded demand, or conversions?

Record the answers against a repeatable prompt set, never off one screenshots. AI-generated responses are variable, and this matters more than most people want it to. A single successful test proves an outcome occurred once. It doesn't prove a stable ranking, and treating it like one is how teams fool themselves.

What Does AI SEO Not Require?

A handful of tactics keep getting sold as requirements for AI visibility even though Google's own documentation says they're unnecessary for Google's generative search features.

[fdb_table title="Common AI SEO claims vs the evidence"]

Claim	What the evidence supports
"You need special AI schema."	Google says no special Schema.org markup is required for AI Mode or AI Overviews.
"You need an llms.txt file for Google AI search."	Google says it ignores llms.txt and other special AI text files for Search.
"Content must be broken into tiny AI chunks."	Google explicitly says chunking is not required.
"Every conversational prompt needs its own page."	Google says systems understand synonyms and meaning; there is no need to rewrite content for every long-tail variation.
"You have to write content with AI."	Content production method and AI-search optimization are different issues. Google focuses on content quality rather than whether AI was used to create it.
"SEO no longer matters."	Google says generative Search remains rooted in its core Search ranking and quality systems.
"A citation or recommendation can be guaranteed."	Eligibility and optimization do not guarantee crawling, indexing, serving or selection. Google explicitly states that serving is not guaranteed.

[/fdb_table]

None of this means every AI-search tactic outside Google's documentation is useless. It means claims need to be scoped to the platform and the evidence behind them. Something can matter for another service without being a Google ranking requirement and a lot of vendors are counting on you not knowing the difference.

Does Structured Data Matter for AI SEO?

Structured data helps describe information explicitly, and it stays valuable for the search features that support it. It's not a special AI-search requirement.

Google recommends that structured data accurately match visible page content, and says there's no additional schema.org markup publishers need to add specifically for AI Mode or AI Overviews. Treat it as one component of a clear information architecture — never as a shortcut around weak content. If your location, product, or service information is unclear on the page itself, markup won't create the missing substance.

Does AI-Generated Content Work for AI SEO?

AI-generated content can perform in search when it's useful and high quality. Using AI carries no special advantage for AI SEO on its own. Google's longstanding guidance cares about the quality and purpose of content, not whether a human or a machine produced it. Using automation primarily to manipulate rankings violates its spam policies.

But AI helping create information doesn't automatically make that information low quality.

The more useful question: does this content contribute meaningful information? 

A language model can summarize public information in seconds, which means generic summaries are now trivially cheap for everyone to produce. And as production costs fall, original data, genuine expertise, first-hand examples, and non-commodity information become more valuable ways of standing out — not less. 

The cheaper generic content gets, the more the premium on real information grows.

Why Do Original Research and First-Party Data Matter for AI SEO?

Original research matters because it creates information gain: useful facts, measurements or analysis that other sources can't provide without referencing the original work.

Say twenty websites explain how consumers choose mortgage brokers. If all twenty repeat the same public statistics, none of them has contributed anything new. Then a company surveys 20,000 mortgage customers and publishes the reasons they picked their lender. That company just created a new evidence source — one that can support its own pages, feed journalists, anchor industry commentary, and end up inside future generated answers.

Original data also gives a brand a legitimate relationship to the topic. It stops merely talking about the subject and starts contributing information to it. Google's guidance toward non-commodity, original, and useful content backs the same distinction.

And this is exactly why FlyDragon's own source context is worth bringing in here. Rather than theorizing about AI citations in the abstract, we can look at what happened in a commercial vertical where we collected the data ourselves.

What We Learned by Studying AI SEO in Real Estate

Real estate makes a useful field study for AI SEO because search results often involve selecting a real-world entity (a specific agent) rather than answering a factual question.

Everything below comes from FlyDragon's 2026 Real Estate AI Citation Index. The findings describe the real-estate recommendation environment we studied. Don't stretch them into universal ranking factors for every AI platform or every industry because that's not what the data supports.

The study audited 5,004 named real-estate agent recommendations and examined the sources and supporting evidence attached to each one.

AI visibility did not simply mirror offline business success

The top 25% of cited agents captured 40.2% of all named recommendations in the dataset; visibility concentrated in a fairly small group. The number I keep coming back to, though, is this one: 46.9% of the top-producer cohort was recommended less often than the median audited agent.

That doesn't prove sales production has no relationship to AI visibility. It demonstrates something narrower and far more useful: high production alone wasn't enough to guarantee high recommendation visibility in the audited responses. 

A person can be commercially successful while the public information environment around them stays weak, inconsistent, or hard to retrieve.

That's an AI SEO problem. And most of the agents affected have no idea they have it.

Owned websites were a major citation source but not the only one

Portals took 22.2% of citation share in our source-category analysis. Agent-owned websites followed close behind at 19.6%, with Google Business Profiles and reviews at 15.2%, brokerage websites at 11.0%, local media at 10.6% and YouTube at 9.8%.

The principle behind that finding: the website matters enormously because it's the information asset the business controls most directly. 

But it doesn't operate alone. 

An entity gets described simultaneously by its own website, platforms, directories, reviews, media, videos and third-party profiles, which means AI SEO has both an owned-information problem and an information-ecosystem problem, and you have to work on both.

Different questions pulled different evidence

The source mix changed materially depending on the intent of the real-estate prompt.

[fdb_table title="Source mix by real estate prompt type" layout="matrix"]

Prompt type	Portals	Agent sites	Reviews	Video	Local media
Best Agent	34%	12%	18%	5%	7%
Luxury Agent	18%	24%	9%	10%	14%
Neighborhood	14%	25%	16%	15%	17%
First-Time Buyer	29%	13%	21%	5%	6%

[/fdb_table]

Broad "best agent" and first-time-buyer prompts leaned heavily on portals and review sources. Luxury and neighborhood prompts shifted toward agent-owned websites, video, and local media.

For my money, this is the most useful finding in the entire dataset, because it kills the lazy version of AI SEO: "get more citations." 

The relevant evidence changes with the search context. A business trying to get selected for a technical query may need different evidence than one competing on reputation. A product comparison may pull different source types than a definition. A local recommendation may use different evidence than an industry statistic. The query determines what needs to be proven.

Different platforms also showed different source patterns

The Citation Index reported different source distributions across the five AI surfaces in its platform analysis. Perplexity showed a larger third-party component in our data, while the Google AI and Gemini samples had larger portal components.

Don't read that as a permanent weighting formula for those platforms. Read it as evidence that one AI visibility score can't describe a brand's presence everywhere. A company can be highly visible on one system and weak on another, because the available sources, retrieval systems, query interpretations, and responses all differ between them.

Accuracy was a separate problem from visibility

We found that 39.3% of the 5,004 audited recommendations had either no visible citation or only partial supporting evidence. Only 43.6% were categorized as fully verified with current, on-point support. 

The audit also turned up stale brokerage information, unverifiable claims, outdated specialties, and flat-out incorrect information.

Being visible and being represented accurately are two different things, and AI SEO should care about both. A business appearing in a generated answer with an obsolete location, a former employer, or a wrong service is visible, and that visibility is doing commercial damage. 

Accuracy, entity consistency and source freshness belong inside AI SEO measurement, not off in a separate housekeeping pile nobody checks.

What Does AI SEO Look Like in Practice?

AI SEO looks different depending on the entity, industry, and search intent being optimized. FlyDragon specializes in residential real estate, so our client work shows how the principles above translate into a live commercial search environment.

These are client case studies, not controlled experiments. Their results shouldn't be treated as guaranteed outcomes for another company or market.

[fdb_table title="FlyDragon AI SEO case studies"]

Case	AI SEO principle demonstrated	Reported outcome
Nate Clark — Austin	Building a clear relationship between an entity and a genuine specialist topic: Nate Clark → probate real estate → Austin	FlyDragon's case study reports Nate becoming the top probate-realtor result across multiple AI platforms, followed by an inbound prospect who said she found him by asking ChatGPT for the best probate realtor.
Richard Berman — Reno	Recommendation visibility + supporting trust across a local commercial query network	Richard's case study reports roughly two AI-sourced listing opportunities per month, a first inbound within two weeks and a highest listing taken of $1.6 million.
Ben Lang — Michigan	AI search as part of an independent research journey before a prospect contacts the business	Ben's case study reports his first inbound in roughly 30 days, four to five high-intent contacts in the first 90 days and one approximately $890,000 transaction with estimated $53.4K GCI before brokerage splits.
April Aberle — Galveston	AI visibility influencing trust before the sales appointment rather than simply generating a click	April's case study reports weekly seller enquiries, two active AI-sourced buyers and prospects arriving at listing appointments having already researched her.

[/fdb_table]

The individual tactics differ from client to client, but the underlying pattern matches the broader definition of AI SEO every time.

How Do You Measure AI SEO?

Measure AI SEO with a combination of search visibility, AI visibility, source visibility, and commercial outcomes.

Traditional Search data still earns its place, because much AI-search discovery remains connected to search indexes and crawlable web content. Google has also started rolling out dedicated Generative AI performance reports in Search Console — as of June 2026, Google said these were reaching a subset of websites and could show impressions, pages, countries, devices and performance over time for generative AI features. 

ChatGPT search referrals can be measured too: OpenAI says referral URLs from ChatGPT Search include utm_source=chatgpt.com, so publishers that permit OAI-SearchBot can spot the inbound traffic in their analytics.

But clicks alone won't capture every AI-search outcome. 

Here's the fuller picture:

[fdb_table title="How to measure AI SEO"]

Measurement	What it tells you
Traditional organic rankings	Whether underlying search visibility is improving
Search impressions	Whether content is being retrieved/displayed more often
AI citations	Whether a page is being used as visible supporting evidence
Brand/entity mentions	Whether the brand appears inside relevant generated answers
Recommendation share	How frequently the entity appears within a defined set of recommendation prompts
Source mix	Which sources are supporting visibility
Cross-platform consistency	Whether visibility exists beyond one AI surface
Accuracy	Whether generated claims about the entity are correct and current
AI referral traffic	Whether users click from AI search to the website
Branded search	Whether research may be creating subsequent brand demand
Leads and conversions	Whether AI visibility is producing commercial outcomes

[/fdb_table]

No single metric deserves to be called the AI ranking. A useful measurement framework tracks the same relevant query network repeatedly and watches what changes.

How Do You Start an AI SEO Strategy?

Start with the information architecture, not with publishing. A sensible implementation runs in this order:

1. Define the entity and search context. Work out what the company, person or product should legitimately be associated with, who searches for it and which questions matter commercially.
2. Measure the current search and AI baseline. Record traditional rankings, important AI prompts, mentions, citations, supporting sources and existing inaccuracies before you change anything.
3. Fix technical discovery problems. Confirm the important information can be crawled, indexed and accessed by the search systems that matter to the business.
4. Map the query network to canonical pages. Decide which questions belong together, which context deserves its own page, and how the pages connect.
5. Improve the information itself. Add clearer definitions, comparisons, evidence, unique expertise, original data, appropriate media and genuinely useful answers where competitors remain incomplete.
6. Strengthen and reconcile external evidence. Correct inconsistent entity information and build legitimate third-party corroboration where the search context calls for it.
7. Measure, refresh and expand. Repeat the original query set, review which sources are appearing, spot new search journeys and improve the network from observed results rather than assumptions.

This order exists to prevent one specific mistake I see over and over: producing enormous quantities of content before knowing what information the search environment lacks.

[fdb_faq title="Common questions"]

[fdb_faq_item question="Can AI SEO Replace Traditional SEO?"]

No. Search-engine discovery, indexing, relevance, and quality remain foundational to the major AI-powered search experiences, so there's nothing to replace. Google explicitly describes its generative Search features as rooted in core Search ranking and quality systems.

[/fdb_faq_item]

[fdb_faq_item question="Is AI SEO the Same as GEO?"]

They overlap substantially, and there is no standardized relationship between the terms. GEO usually refers specifically to visibility within generated answers, while AI SEO gets used more broadly for search optimization across AI-powered search experiences. Google acknowledges GEO as common industry terminology but regards optimization for its own generative Search features as SEO.

[/fdb_faq_item]

[fdb_faq_item question="Is AI SEO the Same as AEO?"]

They overlap too, but AEO focuses more narrowly on becoming a useful answer to a question.

Answer Engine Optimization predates the current wave of generative AI — historically it covered direct-answer surfaces like featured snippets and voice answers. The modern use of AEO increasingly includes conversational and AI-generated answers as well. Search Engine Land similarly treats AI SEO as the broader discipline containing several related optimization concepts.

[/fdb_faq_item]

[fdb_faq_item question="Does AI SEO Work for ChatGPT?"]

Websites can appear in ChatGPT Search, but no amount of optimization can guarantee that ChatGPT will cite or recommend a particular site for a particular prompt.

OpenAI says any public website can appear in ChatGPT Search and recommends allowing OAI-SearchBot when publishers want content discoverable and included in summaries and snippets. Eligibility and selection are two different things.

[/fdb_faq_item]

[fdb_faq_item question="Do You Need llms.txt for AI SEO?"]

You don't need an llms.txt file to improve visibility in Google Search or Google's generative AI features.

Google's current documentation explicitly says it ignores llms.txt for Search, and that creating the file neither helps nor harms Google visibility. Other services can implement different standards, so the correct technical approach is always platform-specific.

[/fdb_faq_item]

[fdb_faq_item question="Do You Need a Separate Page for Every AI Prompt?"]

No. A website doesn't need a separate URL for every conversational phrasing of the same search need.

Google says publishers don't need to capture every long-tail wording or rewrite content specifically for generative search, because its systems understand synonyms and general meaning. Create a separate page when the context, intent or information need genuinely deserves its own document.

[/fdb_faq_item]

[fdb_faq_item question="Can AI SEO Guarantee a Brand Will Be Recommended?"]

No responsible AI SEO strategy can guarantee that an independent AI platform will recommend a particular brand, product or professional. Anyone promising otherwise is selling you something.

Generated responses vary by platform, query, user context, available information and time. Even meeting Google's technical and content requirements doesn't guarantee a page will be crawled, indexed or served. The job of AI SEO is improving the quality and availability of the information those decisions get made from.

[/fdb_faq_item]

[fdb_faq_item question="How Long Does AI SEO Take?"]

There's no universal AI SEO timeline, and you should be suspicious of anyone quoting one.

A newly published page may be discovered quickly. Building broader entity associations, authority, external evidence and stable visibility can take considerably longer. Competition, existing authority, crawl frequency, information quality, platform and query type all move the result.

FlyDragon's client case studies contain examples of fairly fast results, including first inbound opportunities within weeks in individual real-estate campaigns. Those examples show what happened for those clients. They're not a reliable universal timetable, and we don't present them as one.

[/fdb_faq_item]

[/fdb_faq]

What Is the Most Important Part of AI SEO?

The most important part of AI SEO is creating the best possible information environment for the search needs you legitimately deserve to satisfy.

Technical optimization makes information accessible. Semantic structure establishes what it means. Topical coverage establishes the questions it answers. Original research gives the source unique value, independent evidence corroborates the important claims, accurate information keeps the entity consistent, and measurement reveals whether search systems are using any of it. None of those elements works well in isolation.

AI SEO is no hack for inserting a company into ChatGPT. 

It's an extension of the work SEO has always done at its best: helping search systems retrieve the right information from the right source for the right need — now across search experiences that increasingly generate answers instead of only ranking pages.

How FlyDragon Applies AI SEO to Real Estate

FlyDragon specializes in applying these principles to residential real estate, where the entity being retrieved is often the agent themselves.

Our 2026 Real Estate AI Citation Index found substantial differences in which agents were surfaced, which source types supported them, how source mixes changed with query intent and how frequently recommendations were backed by incomplete or outdated evidence.

That research is why our approach starts with one question (what does the search system currently know about this agent, and what evidence does it have to support that understanding?) rather than "how many AI-written blogs can we publish?"

If you're a real estate agent, you can use FlyDragon's AI Visibility Scorecard to see how your own name is currently represented across the search journeys that matter in your market.

AI search optimization starts with something much more basic than most of the tactics people talk about. The system has to be able to access your source, retrieve it for the right question, use it as evidence, and understand which entity that information belongs to. Only then does it make sense to worry about whether the page gets cited, the brand gets mentioned, or the company gets recommended.

There also isn't one universal position to win. ChatGPT, Google AI Overviews, AI Mode, Perplexity, and Copilot all assemble answers around a prompt, a conversation, a location, and a particular moment in time. You can improve your chances of appearing. You can't control the answer.

Ranking in AI Search Has More Than One Outcome

I run AI-search campaigns at FlyDragon, mostly for real estate businesses, and I keep seeing the same mistake.

"Ranking in AI search" gets used as though it describes one thing. It doesn't.

A page can be retrieved, used, cited, mentioned, recommended, or visited. Sometimes several of those happen at once. Sometimes they don't.

Traditional search makes this easier to see because you can point to a URL sitting in a particular position. AI answers are messier. A system might read twenty pages, cite four, mention three companies, and recommend one. You might see traffic from a cited page while another page on your site contributed a fact to the answer without ever receiving a visible link.

Outcome	What happened	What you can usually observe
Retrieved or fetched	The page entered the candidate source pool.	Partial evidence from server logs or platform tools.
Used	The answer drew language, structure, evidence, or facts from the page.	Often an inference from close source comparison.
Cited	A visible source link points to the page.	The citation and its surrounding claim.
Mentioned	The brand, product, or person appears by name.	The captured answer.
Recommended	The entity is selected for the user's task and given a reason.	The wording, order, conditions, and competitors named.
Visited or converted	The user continues to the site or business.	Analytics, calls, forms, sales, or another conversion record.

For reporting, I keep citation rate, mention rate, recommendation rate, and referral conversions separate. You can roll them into one AI visibility score if you want a simple dashboard number, but the trade-off is obvious: once everything is blended together, you lose the explanation for why performance moved.

I expect serious AI-search reporting to start splitting these outcomes out by default over the next twelve months. A single score is useful for a quick glance. It's pretty poor for deciding what to fix.

How AI Search Systems Find and Choose Sources

Different AI-search products use different systems, so there isn't one pipeline we can point to and say, "this is how all of them work." A useful working model is still possible though.

A request gets interpreted. Searches may be issued or rewritten. Candidate sources are retrieved, merged, filtered, or reranked for the current context. Supporting passages are pulled out. The answer is generated, and citations are attached where the product provides them.

I use that sequence as a diagnostic model, not as a claim about a published algorithm.

Google documents query fan-out for its AI features, where a complex request can produce several related searches across subtopics and different data sources. A Google patent for stateful chat describes synthetic queries being built from the user's request and earlier conversation. Another filing describes initial passage retrieval followed by cross-encoder reranking and grounding-span selection.

Patents are useful because they expose possible architectures. They do not prove that a live product uses every step or signal described in them today.

What they do help show is that retrieval and final source selection can happen at different stages. That distinction matters a lot more than people think.

A blocked URL can lose before retrieval. A vague page can be accessible and indexed but still fail to match the search being generated. Another page might match perfectly and then lose during reranking because its evidence isn't strong enough. A company can even earn citations and still fail to get recommended because the broader evidence around that entity is weak.

Find the first stage where the source drops out. That's usually where the work belongs.

Start With a Query Network, Not One AI Keyword

Repeating the same phrase across a title, a few headings, and the body is a very shallow way to think about relevance.

A page becomes useful for a task when it covers the main question and the connected questions that have to be answered before the system can finish the job.

Take:

"What is the best project-management software for a remote design team?"

A decent answer may require information about design approvals, ten-person pricing, Figma integrations, Slack integrations, security, customer reviews, and direct product comparisons. The original query is one part of the problem. Those supporting searches are the rest of it.

That's the query network.

Build it in a spreadsheet with five columns:

• the question or likely search;
• the reader's intent;
• the entity and attribute being requested;
• the best format for the answer;
• the URL that owns it.

The trick is deciding where the intent changes.

One page can own the representative task and the follow-up questions required to answer it. A full comparison, pricing calculator, or implementation manual probably deserves its own URL because the job has changed.

That stops you falling into either extreme: one thin URL for every wording variation, or one gigantic guide trying to cover an entire subject whether the sections belong together or not.

Make the Right Page Eligible to Be Found

Before you get into content quality or citations, the relevant search system has to be able to access the canonical page, interpret its main content, and include it in whatever source pool that product uses.

That's only eligibility. It doesn't mean the page will be selected.

I would check the technical path in this order:

1. Return a successful status for the canonical URL and redirect duplicate versions.
2. Remove accidental noindex, nosnippet, robots, authentication, or firewall blocks.
3. Render the main answer in accessible HTML rather than hiding it inside an image or an interaction that never loads for a crawler.
4. Use a self-referencing canonical and include the URL in an accurate XML sitemap.
5. Link to the page from a relevant hub or guide that search systems already know.
6. Inspect index reports, server logs, and crawler documentation for the surface you care about.

Google says a page needs to be indexed and eligible to show a normal Search snippet before it can appear as a supporting link in AI Overviews or AI Mode. Its newer generative AI search guide also makes a few things very clear: Google doesn't need special AI markup, doesn't use llms.txt for Search, and doesn't require publishers to chop articles into tiny artificial chunks.

OpenAI's publisher guidance gives its crawlers different jobs. OAI-SearchBot is used for ChatGPT search visibility. GPTBot controls potential model-training use. They're separate controls and should be treated that way.

Perplexity documents a similar distinction between PerplexityBot and Perplexity-User.

I wouldn't rely on a universal "AI crawler checklist" for this stuff. Read the documentation for the product you care about. These controls change too often.

Give Every URL One Clear Job

I want to be able to look at the top of a page and answer three questions without doing much work.

What entity is this about?

What job is this page supposed to complete?

Why should I trust this particular source on that subject?

For this article, the central entity is AI-search visibility. The page is meant to teach a site owner how to improve it. The source context comes from FlyDragon's work with semantic SEO, retrieval research, and measured AI-search campaigns.

That's enough.

A 1,500-word detour into the history of large language models might technically be related to the subject, but it would make this page worse at its actual job.

This is where the language you use starts to matter as well.

"OAI-SearchBot supports search discovery" gives us an entity, an action, and a fairly clear purpose.

"OAI-SearchBot is important for AI SEO" doesn't tell us much. "Important" is doing all the work, and it can't really be checked or tied to a specific question without extra interpretation.

I also pay attention to the order of the information. Explain what the outcome is before explaining the mechanism. Explain how the mechanism works before prescribing changes to the page. Once the page work is clear, measurement and diagnosis make more sense.

The page should feel like one argument developing, not a pile of individually optimized sections.

Google's guidance also pushes against creating a fresh page for every long-tail variation. Search systems can connect synonyms and related meanings. Cover the question chain properly, and create another URL when the user is trying to do something different.

Publish Information Worth Retrieving

The best reason for a system to use your page is that the page contains something it needs.

That could be a fact, a method, a comparison, original data, a controlled test, or a useful firsthand observation that isn't available in every other summary on the web.

Formatting makes good information easier to move around. It doesn't make weak information useful.

The original Generative Engine Optimization study reported visibility gains of up to 40% for some methods inside its controlled benchmark. That's interesting research, but I wouldn't turn it into a promise that applying the same techniques today will give you a 40% lift across ChatGPT, Google, or Perplexity. The experiment doesn't support that claim.

A 2026 paper on citation selection and citation absorption looked at 602 controlled prompts, 21,143 valid citations, 18,151 fetched pages, and 72 page features. Pages that had more influence on generated answers tended to be longer, better structured, more semantically matched to the request, and richer in things like definitions, numbers, comparisons, and steps.

Again, correlation inside a designed study isn't a public weighting formula. I treat those features as clues about what makes a source useful enough to retrieve and absorb.

My read from all of this is that structure starts paying off once there's something worth structuring.

When you're adding evidence, give it enough context that someone else can understand what the number or claim means. Name the source. Include the date or method where it matters. Include the limitation when leaving it out would change the interpretation.

Roughly, I would value the evidence like this:

1. First-party data with the collection method, date, and sample.
2. A controlled test or case study with a baseline, intervention, and result.
3. Named experience from a person whose role and scope are visible.
4. Primary documents, standards, and platform guidance.
5. A defensible synthesis that connects those sources in a new way.

A statistic roundup is useful when the page exists to collect statistics. We already maintain FlyDragon's current AI SEO statistics for that.

On an implementation page, I use a simpler test: does this number change what the reader should do? If it doesn't, I probably don't need it.

Make Each Answer Easy to Extract

I don't think every page needs to be broken into fifty miniature answers for AI.

What matters is whether a section answers the heading and whether that answer still makes sense when it is pulled away from the rest of the article.

Usually that means answering the heading fairly early, then choosing a format that suits the information. Processes work well as numbered steps because order matters. Repeated comparisons usually belong in a table. A set can use bullets. Explanations normally read better as prose.

The content decides the format.

Definitions should keep the conditions that make them true. Processes should retain their order. Comparisons should use consistent fields. If you're making a data claim, keep the source, date, method, or limitation close enough to the number that someone can't accidentally lift the exciting part and leave the caveat behind.

Basic publishing choices help here too. Use headings that describe the question or decision. Give tables actual column headers. Write anchor text that tells the reader where they're going. Keep the important answer in visible HTML.

Just don't over-engineer it.

A page made from sentence fragments, hundreds of micro-sections, forced FAQ blocks, and the same phrase repeated over and over might look "optimized" in a spreadsheet. It usually reads terribly.

Google says there is no ideal page length and no requirement for artificial chunking. Three thousand useful words are better than five thousand words written to hit some arbitrary content target.

Build Corroboration Beyond Your Website

Your own website tells the system what you say about yourself. Independent sources help establish whether those claims line up with the rest of the web.

That can cover the company, person, product, location, reputation, category, services, or whatever else the system needs to understand.

I want basic entity facts to be boringly consistent across good sources. Names, locations, descriptions, authors, products, categories. If an important third-party profile has the wrong location or describes a service you stopped offering three years ago, fix it.

Beyond that, real corroboration comes from things like editorial coverage, expert quotes, customer reviews, good reference listings, and other evidence you've earned.

This starts mattering even more when the query moves from "tell me about this" to "which one should I choose?"

Your own guide might be a perfectly good source for explaining how project-management software works. If the user wants the best option for a regulated enterprise buyer, the answer may also need security documentation, independent reviews, customer evidence, and accurate product information before it can confidently recommend anything.

Fake reviews, paid forum spam, mass-produced mentions, and invented consensus are a different thing entirely. They leave a dirty evidence trail, and Google's current guidance already tells site owners to ignore inauthentic mention schemes.

Build evidence you'd be happy to show somebody.

Use Internal Links as Semantic Bridges

I mostly think about internal links as a way to explain relationships.

They help people find the next useful page, but they also show search systems how one topic or task connects to another.

For this subject, a sensible structure might look like:

AI SEO hub → ranking implementation guide → visibility audit, statistics evidence, and platform update pages.

Supporting pages can link back to the parent process where the relationship is useful. The anchor should describe what the destination page does. It doesn't need to repeat the same exact keyword every time.

Context is much more useful than hitting some arbitrary internal-link quota.

If I'm talking about measurement and link to a guide explaining how to measure AI visibility, the reason for the link is obvious. If I dump fifty vaguely related articles into a footer, I've technically created more internal links without explaining much of anything.

I also remove self-links, duplicate anchors where they serve no purpose, and links added purely because somebody thinks they need to pass "SEO juice."

Google, ChatGPT, Perplexity, and Bing Expose Different Controls

Most of the durable page work overlaps across the major AI-search products. The product-specific stuff changes much faster.

Access controls change. Source indexes are different. Citation interfaces are different. Publisher reporting is different.

So I keep that layer separate from the main methodology.

Surface	Documented access or eligibility	First place to look	Measurement available to publishers
Google AI Overviews and AI Mode	Google Search indexing and snippet eligibility.	Technical SEO, page usefulness, and relevant internal links.	Google Search Console's generative AI reporting where available.
ChatGPT Search	OAI-SearchBot for search discovery; GPTBot is a separate training control.	Intended crawler access and source-backed pages that web search can retrieve.	Referral analytics plus repeated prompt tracking.
Perplexity	PerplexityBot for indexing and Perplexity-User for user-requested fetches.	Robots and firewall access, stable canonical URLs, and useful evidence.	Captured answers, citations, referrals, and server evidence.
Bing and the Copilot ecosystem	Bing crawl and index systems.	Bing indexability and page-level citation evidence.	Bing Webmaster Tools AI Performance reports citations, cited pages, and grounding queries.

Bing says its AI Performance report does not tell publishers a citation's importance, placement, or rank. That's an important limitation. Seeing ten citations tells you your pages made it into ten answers. It doesn't tell you whether those answers leaned heavily on your source, mentioned it in passing, or endorsed the business.

For the Google-specific changes, we keep a separate review of Google's latest AI-search guidance. The implementation system on this page doesn't need rewriting every time one platform changes a report or crawler rule.

Measure AI Visibility as a Rate

A single AI answer is one observation. I've seen people put far too much weight on one run.

If you want useful measurement, repeat a controlled set of prompts across products, runs, and dates, then record the outcomes separately.

For every run, save the prompt, platform, product surface, account state, location, date, answer, cited URLs, brands named, and recommendation order.

If you change something on the site, keep the original prompt set intact. Add new questions to a second cohort rather than changing the test halfway through and then comparing the results as though the baseline stayed the same.

The basic citation-rate calculation is:

Citation rate = eligible answer runs with at least one citation to the tracked domain ÷ all eligible answer runs × 100

The paper Don't Measure Once documents variation across repeated runs, prompt wording, and time, which is exactly why I don't like one-off screenshots being treated as measurement.

It doesn't give us one magic sample size either. The number of runs you need depends on how much variance you're seeing, what decision you're trying to make, and how much it costs to collect another answer.

We've got a full process for a repeatable AI visibility audit if you want the prompt design, scoring sheet, and comparison method.

For the purpose of this guide, keep the experiment boring: compare the same things under the same conditions and save the raw answers.

Diagnose the Earliest Failed Stage

This is probably the most useful distinction in the whole process.

A crawl problem needs a different fix from a retrieval problem. A retrieval problem needs a different fix from a citation-selection problem. Being cited and never recommended is another problem again.

Symptom	Likely stage	First evidence to inspect	First repair
The page is absent from the relevant search index.	Access or eligibility.	Index report, robots, canonical, response, and rendering.	Remove the specific technical blocker.
The page is indexed but never appears for the topic.	Demand, relevance, or coverage.	Prompt set, competing sources, page intent, and entity relationships.	Repair intent mapping and the missing question chain.
The page is fetched or surfaced but receives no citation.	Reranking, evidence, or extraction.	The competing passages and their supporting proof.	Strengthen the exact answer and the evidence behind it.
The page is cited but contributes little to the answer.	Use or absorption.	Distinctive facts, language, and structure carried into the answer.	Add material the synthesis needs and can attribute.
The brand is mentioned but never recommended.	Entity fit or corroboration.	Third-party sources, reviews, category fit, and proposition clarity.	Improve verifiable off-site evidence.
The result changes sharply between runs.	Context or measurement.	Prompt wording, prior conversation, account, location, product, and date.	Repeat controlled runs and report the distribution.

The easy mistake is letting each team fix the problem they already know how to fix.

Writers rewrite copy. Developers check robots.txt. PR teams go looking for more mentions.

Sometimes that's the right answer. Sometimes it has nothing to do with the stage that's failing.

Follow the evidence first.

Your First 30 Days of AI-Search Work

Thirty days is plenty of time to fix the obvious eligibility and intent problems, publish at least one useful evidence-led improvement, and establish a baseline you can repeat.

It is not a promise that every search product will recrawl the page, select it, and start citing it inside thirty days. Those are different things.

1. During days 1–5, define the tracked questions, record current outcomes, assign each intent to one URL, and find pages competing with each other.
2. During days 6–10, repair indexing, crawler access, canonicals, rendering, sitemaps, and internal discovery.
3. During days 11–20, tighten the central entity and source context, close the necessary question chain, add original evidence, and use the right format for each answer.
4. During days 21–30, correct third-party facts, pursue legitimate corroboration, repeat the baseline conditions, and log which stage moved.

If you're hiring somebody else to do this, use these questions to ask an AI SEO agency before buying guaranteed citations or a mystery visibility score nobody can explain.

Pages rarely win because somebody discovered one magic markup trick.

They answer a real question, survive the stages that happen before synthesis, and give the system better evidence than the alternatives. Then you measure what happened and work on whichever part broke first.

Frequently Asked Questions

Can you guarantee a ranking in AI search?

No. You can improve access, relevance, evidence, extraction, and corroboration. The search product still controls its own retrieval, generation, and citations.

If somebody guarantees a specific AI ranking or citation outcome, I'd want to see very strong evidence for how they're making that promise.

Do I need llms.txt to rank in AI search?

Google says it does not use llms.txt for Search, including its generative AI features.

Another service may decide to use the file, so check its documentation. I wouldn't treat llms.txt as a universal AI-search requirement.

Does schema markup make a page rank in ChatGPT or AI Overviews?

There is no documented source supporting that guarantee.

Use structured data where it accurately describes the entity or rich-result type it was designed for. It doesn't replace visible content, crawler access, relevance, or evidence.

Can a page appear in an AI answer without ranking number one on Google?

Yes.

AI features can retrieve and rerank sources for different parts of the question, so a page doesn't need to hold a universal number-one Google position before it can appear in an AI answer.

That doesn't make normal search visibility irrelevant. Search-grounded products still rely heavily on crawlable, indexable sources.

How long does it take to rank in AI search?

There isn't one honest universal timeline.

Crawl and index refreshes, competition, the quality of the source, product behavior, off-site corroboration, and how often you're measuring can all change how quickly a result becomes visible.

Is AI-search optimization different from SEO?

I see it as an extension of SEO rather than a completely separate discipline.

The technical SEO, useful content, internal linking, and authority work still matter. AI search optimization adds another layer around query fan-out, retrieval, passage selection, answer synthesis, citations, mentions, recommendations, and repeated measurement.

How can I tell whether an AI answer used my page without citing it?

Sometimes you can't prove it.

Keep going: the AI search playbook

• LLM SEO, explained in full
• What AI-powered search is
• The best AI search engines for leads
• Writing a bio AI search can use
• How AI search changed the way people find providers
• Using video to rank in AI search
• Generating leads from ChatGPT
• What the Zillow-ChatGPT partnership means
• ChatGPT ads, explained
• Will AI replace real estate agents?
• How generative AI is changing real estate
• Why AI content is the new spam

And if you'd rather someone did this work for you — that's the job. Our GEO service exists because most businesses have the evidence and nobody's ever assembled it.