SEO, AEO and GEO Guide

Search is changing. People still type queries into Google and Bing, but they also get direct answers in featured snippets, voice assistants and AI-generated summaries, and they ask AI assistants questions that used to start with a search. Three terms have emerged to describe the work of being visible in all of these places: SEO, AEO and GEO.

This guide explains what each term means, how much they overlap, and what you can do in practice. It is deliberately honest about uncertainty. SEO is a mature discipline with official documentation from search engines. AEO and GEO are emerging practices, and nobody (not agencies, not tool vendors, not us) controls whether an AI system cites a particular page.

Who this guide is for and what you will be able to do

This guide is for small-business owners, freelancers, content creators and marketers who manage a website and want more people to find it through search and AI tools. It starts from the basics and builds up to intermediate topics such as structured data, crawler controls and content structure for answers.

By the end you will be able to:

• Explain SEO, AEO and GEO clearly and know where they overlap and differ.
• Run a basic technical health check on your site.
• Research what your audience searches for and plan content around it.
• Write pages that are easy for search engines, answer engines and AI systems to understand and quote.
• Use structured data appropriately, without expecting it to work miracles.
• Decide how to handle AI crawlers in your robots.txt file.
• Measure progress realistically and avoid common mistakes.

Search engines and AI products change frequently. Where something depends on a current product or policy, we say "at the time of writing". Check official documentation, especially Google Search Central, Bing Webmaster guidelines and each AI operator's crawler documentation, before relying on specifics.

SEO, AEO and GEO defined

SEO: search engine optimisation

SEO is the practice of making a website easier for search engines to crawl, understand and rank, and more useful to the people who arrive from search results. It covers technical foundations (can search engines access and index your pages?), content (does the page satisfy the searcher's need?), and signals of trust and authority (do other reputable sites reference you? do you demonstrate real expertise?).

The main outcome of SEO is appearing in organic (unpaid) search results and earning clicks.

AEO: answer engine optimisation

AEO is the practice of structuring content so that systems which give direct answers can find, extract and present it. "Answer engines" originally referred to featured snippets, "People also ask" boxes, knowledge panels and voice assistants that read out a single answer. The term is now also used for AI answers.

AEO emphasises clear questions and concise answers, logical headings, definitions, lists, tables and accurate structured data. The outcome is being shown as the answer, sometimes with a click to your site and sometimes without one.

GEO: generative engine optimisation

GEO is the practice of increasing the chance that generative AI systems (such as AI summaries in search, and AI assistants and chatbots that search the web) mention, summarise or cite your content or brand. These systems generate a written answer, often drawing on multiple sources, and may show links to some of them.

GEO emphasises being accessible to the relevant crawlers, being a clear and credible source on a topic, being mentioned by other reputable sources, and writing content with specific, verifiable facts that are easy to quote accurately. Because each AI product works differently and changes often, GEO is the least predictable of the three.

Where they overlap

The honest summary: good SEO is the foundation of both AEO and GEO. Google says, at the time of writing, that there are no additional requirements or special optimisations needed to appear in its AI Overviews or AI Mode beyond its normal Search guidance, and that you do not need special files or schema for them. Most of what people call AEO and GEO is good SEO and good writing, with extra attention to clarity, structure, factual precision and crawler access.

What nobody can promise

No one can guarantee rankings, featured snippets or AI citations. Search engines and AI providers decide what to show based on their own systems, which consider many factors and change frequently. Be cautious of anyone selling guaranteed AI visibility. What you can do is remove barriers, publish genuinely useful content, and make it easy to understand and quote.

How search engines and AI systems find content

The classic search pipeline

1. Crawling: a crawler (such as Googlebot or Bingbot) discovers URLs through links and sitemaps and downloads pages.
2. Rendering: the search engine processes the HTML and, where needed, runs JavaScript to see the final content.
3. Indexing: it analyses the content and stores it in its index.
4. Ranking and serving: when someone searches, it selects and orders results from the index.

If a page fails at any step, it will not appear. Many SEO problems are simply pages that cannot be crawled or are not indexed.

How AI answers draw on the web

AI systems use web content in two broad ways:

• Training: some crawlers collect content that may be used to train AI models. Content used in training can influence what a model "knows", but models do not usually cite training sources.
• Retrieval (search and grounding): some AI products search the web or their own index at the time of the question, then generate an answer from the retrieved pages and may link to them. This is where citations usually come from.

AI features inside search engines (such as Google's AI Overviews) generally draw on that search engine's index. Standalone assistants use their own crawlers, a search partner's index, or both. The practical implication: to be retrievable, your pages need to be crawlable by the relevant systems and indexed where those systems look.

Step 1: technical foundations

Make sure pages can be crawled and indexed

• Check that important pages return a 200 status code and are not blocked in robots.txt.
• Make sure important pages do not carry a noindex meta tag by mistake.
• Use canonical tags to indicate the preferred version of duplicate or near-duplicate pages.
• Submit an XML sitemap listing your indexable pages. The XML sitemap generator can create one for smaller sites.
• Verify your site in Google Search Console and Bing Webmaster Tools, and check indexing reports.

Fix redirects and broken links

Redirect chains and broken internal links waste crawl effort and frustrate visitors. Use the redirect checker and broken link checker to find problems.

Speed and mobile experience

Search engines primarily evaluate the mobile version of your site. Pages should load quickly and be easy to use on a phone. Compress images (the image compressor helps), avoid heavy scripts, and test with the website speed checker.

Make important content visible in the HTML

If key content only appears after complex JavaScript runs, or only after a user clicks, some crawlers may not see it. Google can render JavaScript, but many other crawlers, including some AI crawlers, may not render it as fully. Where possible, include your main content, headings and links in the server-delivered HTML.

Run an audit

The website SEO audit checks common on-page and technical issues such as titles, descriptions, headings, image alt text and indexability signals, so you can prioritise fixes.

Free toolWebsite SEO AuditFetch any page and see its status, redirects, title, meta tags, canonical, headings, links, schema and robots.txt rules.

Technical foundations checklist

• Site verified in Google Search Console and Bing Webmaster Tools
• Important pages return 200 and are indexable (no accidental noindex)
• robots.txt does not block important pages or resources such as CSS and JavaScript
• XML sitemap submitted and up to date
• One canonical URL per piece of content (https, consistent trailing slash and www choice)
• No redirect chains or broken internal links on key pages
• Pages load quickly and work well on mobile
• Main content and navigation present in the HTML
• HTTPS everywhere, with no mixed content warnings
• Custom 404 page that helps visitors find their way

Step 2: understand what people search for

Map topics to intent

Every search has an intent. Matching your page to that intent matters more than including a keyword a certain number of times.

Before writing, search the query yourself and look at what already ranks. If the results are all step-by-step guides, a sales page is unlikely to fit. If an AI summary or featured snippet appears, notice what question it answers and how.

Find the questions behind the keywords

For AEO and GEO, questions matter. Gather them from:

• "People also ask" boxes and related searches
• Questions your customers ask in emails, calls and sales conversations
• Your Search Console queries report (look for queries phrased as questions)
• Forums and communities in your industry
• Your support inbox and FAQ requests

Group questions into topics. A topic with many related questions often deserves a comprehensive guide, with individual questions answered clearly within it.

Prioritise with a simple scoring framework

Not every topic deserves a page right now. Score each candidate from 1 to 5 on three factors:

• Business value: how closely it relates to what you sell.
• Intent fit: how likely a searcher is to take a useful action.
• Ability to win: whether you can realistically create the most useful page (considering your expertise and the strength of what already ranks).

Multiply the three scores for a priority number out of 125.

Worked example (illustrative):

The service page comes first, the pricing guide second. The radiator guide may still be worth writing for helpfulness and brand trust, but it is not the top priority.

Step 3: write content that deserves to be found

Start with people, not algorithms

Google's guidance, at the time of writing, focuses on "helpful, reliable, people-first content". That means content created mainly to help people, written by or with input from someone who knows the subject, and showing first-hand experience where relevant. Pages written mainly to attract search traffic, with little original value, tend to perform poorly over time.

Practical signs of helpful content:

• It answers the question fully, including the follow-up questions a reader would have.
• It includes original information: your own examples, data, photos, processes, prices or opinions based on experience.
• It is clear who wrote it and why they are qualified, with an author or company page.
• It is accurate and kept up to date, with dates where they matter.

Structure pages for answers

This is where AEO and GEO practices add most value. Structure helps readers skim, and it helps systems identify and extract the right passage.

1. Put a direct answer near the top. If the page answers "What is a break-even point?", give a clear one to three sentence definition early, then expand.
2. Use descriptive headings that reflect the questions people ask. "How much does a boiler replacement cost?" is clearer than "Pricing thoughts".
3. Make each section self-contained. A passage that makes sense when quoted on its own is easier for an answer engine to use accurately. Avoid relying on "as mentioned above".
4. Use lists for steps and tables for comparisons. They are easy to parse and quote.
5. Define terms the first time you use them.
6. Be specific. "Most installations take one to two days for a like-for-like swap" is more useful (and more quotable) than "installation is quick". Only state what you can support.
7. Cite sources for facts that are not your own, and link to them.

Example (before): "There are loads of factors that go into what you'll pay and it really depends, so get in touch."

Example (after): "A boiler replacement price depends mainly on three things: the boiler type (combi, system or regular), whether the new boiler goes in the same location, and how much pipework needs changing. Moving a boiler to a new room usually costs more than a like-for-like swap because of the extra labour and materials."

The second version is more useful to a reader and easier for any system to understand and quote.

The AEO content readiness checker reviews a page for answer-friendly structure such as question headings, concise answers, lists and definitions. The AI FAQ generator and AI blog outline generator can help you draft question lists and outlines, which you should then check and improve with real expertise.

Free toolAEO Content Readiness CheckerCheck a page or draft for answer-friendly structure: question headings, concise answers, lists, schema, authorship and dates.

Titles and meta descriptions

The title tag strongly influences what appears as the clickable headline in search results (search engines may rewrite it). Keep titles descriptive, unique for each page and front-loaded with the main topic. Meta descriptions do not directly affect ranking in Google, but a clear description can improve the snippet and encourage clicks, though search engines may generate their own snippet instead.

Use the meta title generator and meta description generator for ideas, and the SERP preview to see roughly how your title and description may appear before publishing. Display lengths depend on pixel width and device, so treat any character guideline as approximate.

Why click-through rate is worth improving

Worked example (illustrative): A page receives 10,000 impressions a month in search with a 2% click-through rate.

• Clicks = 10,000 x 0.02 = 200
• If 2% of visitors enquire: 200 x 0.02 = 4 enquiries

If a clearer title and description lift the click-through rate to 3% with the same position and impressions:

• Clicks = 10,000 x 0.03 = 300
• Enquiries = 300 x 0.02 = 6

That is 50% more enquiries from the same ranking. Results vary, and search engines may change how your snippet appears, but titles and descriptions are among the cheapest things to test.

Internal linking

Link related pages to each other with descriptive anchor text. This helps search engines discover pages and understand how topics relate, and helps readers find more detail. A guide on a broad topic should link to the detailed pages on subtopics, and those pages should link back.

Keep content fresh where it matters

Update pages when facts, prices, laws or products change, and show an updated date. Do not change dates without making real updates. For fast-changing topics, schedule reviews.

Step 4: structured data

What structured data does

Structured data (usually in JSON-LD format using the schema.org vocabulary) describes the content of a page in a standard, machine-readable way: this is an article with this author and date, this is a local business with these opening hours, this is a product with this price.

In Google Search, structured data can make pages eligible for rich results (enhanced displays such as review stars, product details or event listings), provided the page meets Google's guidelines. Eligibility is not a guarantee: Google decides whether to show a rich result.

What structured data does not do

At the time of writing, Google states that no special schema.org structured data is needed to appear in its AI features. Structured data may help systems understand your pages, but there is no official evidence that adding it guarantees AI citations. Treat it as good practice for clarity and rich result eligibility, not as a GEO shortcut.

Also note that Google has reduced some rich result types over time. At the time of writing, FAQ rich results are shown only for well-known, authoritative government and health websites, and HowTo rich results are no longer shown. You can still use clear question-and-answer content on your pages; just do not expect FAQ markup to create a visual enhancement for most sites.

Useful types for small businesses

Rules to follow

• Mark up only content that is visible on the page.
• Keep structured data accurate and consistent with the page.
• Do not mark up fake reviews or self-serving ratings that break guidelines.
• Validate before publishing.

Create markup with the schema markup generator and check it with the structured data validator. Google's Rich Results Test is also useful for checking eligibility.

Free toolSchema Markup GeneratorFill in a form to generate JSON-LD for Organization, LocalBusiness, Article, Product, FAQ, Breadcrumb, Event and more.

Step 5: AI crawlers and robots.txt

What robots.txt does and does not do

The robots.txt file at the root of your site (for example https://example.com/robots.txt) tells crawlers which parts of the site they may crawl. Key points:

• It is a voluntary standard. Reputable crawlers follow it; bad actors may ignore it.
• It controls crawling, not indexing. A blocked URL can still appear in search results if other sites link to it. Use noindex (on a crawlable page) to keep a page out of search results.
• It is not a security tool. Never rely on it to hide private content; use authentication instead.
• Blocking a crawler now does not remove content it has already collected.

Common crawler tokens

At the time of writing, these are some commonly discussed user agent tokens, based on their operators' documentation. Names, purposes and behaviour change, so verify with each operator before editing your file.

Notice the pattern: several operators now separate training crawlers from search or retrieval crawlers. That lets you make different decisions for each.

A decision framework

Ask what you want:

1. Maximum visibility in search and AI answers: allow search engines and AI search or retrieval crawlers. Decide separately whether you also allow training crawlers.
2. Visibility in AI search, but not model training: allow search and retrieval crawlers (such as OAI-SearchBot, Claude-SearchBot and PerplexityBot) and disallow training crawlers (such as GPTBot, ClaudeBot, CCBot, and the Google-Extended token). Understand that operators define and implement these categories themselves.
3. Minimum AI use: disallow AI crawlers you do not want. This may reduce your visibility in those AI products. Do not block Googlebot or Bingbot unless you want to disappear from their search results.

Paywalled, licensed or premium content may justify stricter rules. A local service business that wants to be recommended generally benefits from being accessible.

Example robots.txt (illustrative, option 2)

User-agent: *
Allow: /

User-agent: GPTBot
Disallow: /

User-agent: ClaudeBot
Disallow: /

User-agent: CCBot
Disallow: /

User-agent: Google-Extended
Disallow: /

Sitemap: https://example.com/sitemap.xml

This allows all crawlers by default, then disallows specific training-focused tokens. Adapt it to your own decision, and re-check tokens against current documentation. The robots.txt generator builds a file from your choices, and the AI crawler access checker shows which AI crawlers your current file allows or blocks.

Also check your hosting, CDN or firewall settings. Some services block certain bots by default, which can override what your robots.txt allows.

Free toolAI Crawler Access CheckerSee whether GPTBot, ClaudeBot, PerplexityBot, Google-Extended and other AI crawlers may access a URL under your robots.txt.

What about llms.txt?

You may see recommendations to add an llms.txt file, a proposed convention for giving AI systems a curated summary of a site. At the time of writing it is not an official standard, and Google states that you do not need special AI text files to appear in its AI features. Adding one is unlikely to hurt, but do not expect it to change visibility, and prioritise the fundamentals first.

Step 6: authority, mentions and reputation

Why others' words about you matter

Search engines have long used links from other sites as signals. AI systems that generate answers also tend to draw on sources that are widely referenced and trusted on a topic. Being mentioned accurately and positively across the web (in industry publications, local news, directories, review platforms and community discussions) supports all three disciplines.

Practical, legitimate ways to earn mentions

• Publish original data, research or tools others want to reference.
• Contribute expert commentary to journalists and industry publications.
• Keep your business listings consistent (name, address, phone number) across Google Business Profile and major directories.
• Encourage genuine customer reviews on relevant platforms, and respond to them.
• Participate helpfully in communities where your audience asks questions, following each community's rules.
• Partner with complementary businesses on useful content.

Avoid buying links or taking part in link schemes. These breach search engine guidelines and can lead to lost rankings.

Make your brand facts consistent

AI systems can repeat inaccurate details if the web contains conflicting information about you. Keep your about page, contact details, services, prices and key facts consistent across your site and profiles. Clear Organization or LocalBusiness structured data can support this.

Step 7: measure progress realistically

What you can measure well

• Search Console: impressions, clicks, click-through rate and average position by query and page, plus indexing status. At the time of writing, Google includes traffic from its AI features within the overall Search performance data rather than as a fully separate report, so check the current documentation.
• Analytics: visits, engagement and conversions from organic search.
• Referrals from AI products: some AI assistants pass referrer information when a user clicks a cited link, so you may see them as referral sources in analytics. Not all do, and some visits may appear as direct traffic.

What is hard to measure

• How often an AI answer mentions your brand without linking.
• How often AI summaries answer a question so that the user does not click at all.

Some third-party tools sample AI answers for a set of prompts to estimate visibility. These can show trends, but results vary by user, location, time and product version, so treat them as indicators, not precise measurements.

A monthly review routine

1. Check indexing issues and fix new errors.
2. Review the top pages and queries by clicks, and those with falling clicks.
3. Look for pages with many impressions but low click-through rate, and test better titles and descriptions.
4. Look for queries where you rank on page two (positions 11 to 20) and improve those pages.
5. Record conversions from organic search and, where visible, AI referrals.
6. Manually search a handful of your most important questions in major AI tools and note whether and how your brand appears. Keep notes over time.

Content review checklist

Use this before publishing or refreshing any important page.

• The page matches the search intent seen in current results
• A direct, accurate answer appears near the top
• Headings describe the questions or topics each section covers
• Each section makes sense if quoted on its own
• Steps are in numbered lists, comparisons in tables
• Facts are specific, current and supported, with sources where needed
• Author or business expertise is clear
• Title and meta description are unique and descriptive
• Internal links point to and from related pages
• Structured data (if used) matches visible content and validates
• Images are compressed and have descriptive alt text
• The page is crawlable by the search and AI crawlers you want to allow

Common mistakes and how to avoid them

Chasing AI hacks instead of fundamentals

Hidden text aimed at AI systems, prompt-like instructions in page content or mass-produced pages do not build durable visibility and can breach search engine spam policies. Focus on crawlability, helpful content and clarity.

Blocking crawlers by accident

A leftover Disallow: / from a staging site, a firewall rule or a CDN bot setting can remove you from search or AI products. Check your robots.txt and server settings after every site launch or migration.

Blocking Googlebot to avoid AI Overviews

Blocking Googlebot removes you from Google Search entirely. Google-Extended does not control AI Overviews in Search. If you want to limit how your content is shown, review Google's preview controls such as nosnippet, understanding they also affect regular snippets.

Publishing thin or generic content at scale

Many near-identical pages (for example, a location page for every town with only the name changed) rarely help users and can hurt the site overall. Create pages only where you have something distinct to say.

Using AI to write without expert review

AI tools can speed up drafting, but they can produce errors and generic text. Always add first-hand knowledge, check facts and edit for clarity. The AI marketing productivity guide covers responsible workflows.

Treating structured data as a ranking trick

Structured data helps understanding and rich result eligibility. Incorrect or misleading markup can lead to manual actions. Keep it accurate.

Ignoring conversion

Traffic without enquiries or sales is not the goal. Make sure pages that attract visitors guide them to a clear next step. See the website growth and conversion guide.

Expecting quick or guaranteed results

SEO usually takes months, and AI visibility is uncertain. Use paid channels, such as those in the Google Ads beginner's guide, when you need traffic sooner.

A 90-day plan

Glossary

AEO (answer engine optimisation): Structuring content so systems that give direct answers can find, extract and present it.

AI Overviews: AI-generated summaries that appear in some Google Search results, drawing on web sources.

Canonical tag: An HTML element indicating the preferred URL for duplicate or similar content.

Crawler (bot, spider): Software that automatically fetches web pages.

Featured snippet: A highlighted answer extracted from a web page, shown at the top of some search results.

GEO (generative engine optimisation): Increasing the chance that generative AI systems mention, summarise or cite your content.

Grounding: When an AI system uses retrieved sources to base its answer on current information.

Indexing: Storing and organising crawled content so it can be shown in search results.

JSON-LD: A format for adding structured data to a page using JavaScript Object Notation.

Meta description: An HTML summary of a page that search engines may show in results.

Noindex: A directive telling search engines not to include a page in their index.

Rich result: An enhanced search result display made possible by structured data.

robots.txt: A file telling crawlers which parts of a site they may crawl.

Schema.org: A shared vocabulary for structured data.

Search intent: The goal behind a search query.

SEO (search engine optimisation): Making a site easier for search engines to crawl, understand and rank, and more useful to searchers.

SERP: Search engine results page.

Title tag: The HTML element that often provides the clickable headline in search results.

User agent: The name a crawler or browser uses to identify itself.

XML sitemap: A file listing the URLs you want search engines to discover.

Next steps

1. Run the website SEO audit and fix crawling and indexing issues first.
2. Check your crawler settings with the AI crawler access checker and decide your policy.
3. Pick your three highest-priority pages and improve them with the content review checklist and the AEO content readiness checker.
4. Improve titles on high-impression pages using the meta title generator and SERP preview.
5. Add accurate structured data with the schema markup generator.
6. Review progress monthly, and combine organic work with paid channels from the Google Ads beginner's guide and Meta Ads and lead generation guide when you need faster results.