# SEO in the AI Era: How to Conquer Search and the Answer Engines

SEO · Published 2026-11-12 · Updated 2026-11-12 · RankMonks
Source: https://rankmonks.com/blog/seo-in-the-ai-era

> Short answer: SEO in the AI era is the work of being found in search results and named in AI answers. The foundations are the same: a site machines can read, pages that answer real questions, consistent facts and a reputation other sites confirm. What changes is how you write, what you mark up and what you measure.

SEO in the AI era is the work of making your business easy to find in search results and easy to name in AI answers. The two are not separate jobs. The same pages, facts and reputation feed both, and a business that neglects one usually slips in the other.

What has changed is where the decision gets made. A buyer used to scan ten links and pick one. Now they may read a summary at the top of Google, ask ChatGPT for a shortlist, or let an assistant compare three options and report back. In each case a machine reads the web first and decides who gets mentioned.

This guide explains how that works, gives you one playbook that covers search and answer engines together, and then shows how seven kinds of business can put it to use: estate agents, small restaurants, local trades, law firms, clinics, online shops and software companies.

## Key takeaways

- AI answers are built on the open web, so the pages and facts you publish, and what others publish about you, still decide whether you appear.

- Search engine optimization, answer engine optimization and generative engine optimization share the same foundations and work best as one programme.

- Answer engines favour passages that answer one question clearly and can be quoted without the rest of the page.

- Small and local businesses have a real opening, because specific first-hand detail is exactly what generic content lacks.

- Consistent business facts across your site, your listings and third-party sites matter more than before, because machines cross-check them.

- Measurement now has two layers: classic search data, and a regular check of what AI tools say when asked the questions your buyers ask.

## What has changed about search in the AI era?

Three things have changed: where answers appear, how questions are phrased and who does the reading. The underlying need, finding a trustworthy answer or supplier, has not changed at all.

[](/posts/seo-in-the-ai-era/then-and-now.svg)

**Answers now appear before the links.** Google shows AI Overviews above the normal results for many questions, alongside older features such as featured snippets and People Also Ask. Many searches now end without a visit to any site, a pattern covered in our guide to [zero-click searches](/blog/zero-click-searches).

**Questions have become longer and more specific.** Someone typing into a search box might write "estate agent Leeds". Someone talking to an assistant writes "which estate agents in north Leeds are good with first-time sellers and what do they charge". The second question contains needs that a generic page never addresses.

**Machines read on the buyer's behalf.** Tools such as ChatGPT, Perplexity, Gemini and Claude can search the web, read several pages and write one combined answer. The buyer may never see your page, only what the model took from it.

There are three terms for the work that follows from this.

| Term | What it aims for | Where the result shows |
| --- | --- | --- |
| SEO, search engine optimization | Your pages rank | Classic search results and map results |
| AEO, answer engine optimization | Your content is used as the direct answer | Featured snippets, AI Overviews, voice assistants |
| GEO, generative engine optimization | Your brand is mentioned and cited | ChatGPT, Perplexity, Gemini, Claude |

The labels are useful for planning, but the work overlaps heavily. That is the main idea of this guide: build the foundations once, and they pay off in all three places.

## Is SEO still worth doing when AI answers the question?

Yes. AI answer engines depend on the same web that search engines index, so the work that makes you visible in search is the work that makes you available to AI.

Consider how an assistant answers a question about local services or current products. It cannot rely on memory alone, because prices, opening hours and stock change. It runs searches, reads the pages that come back and writes from them. If your page is not crawlable, not indexed or not clear, it is not in the pile the model reads from.

Search itself has not gone away either. People still search when they want to visit a specific site, compare options in detail, buy something or find a place on a map. These are the searches closest to a sale, and they still produce clicks.

What has declined is the value of thin informational pages. A page that answers "what is conveyancing" in the same words as a hundred others has little to offer when the definition appears on the results page. The pages that keep their value are the ones with something a summary cannot replace: your prices, your process, your local knowledge, your examples and your evidence.

So the honest position is this. SEO is still worth doing, but the target has shifted from "get the click on every query" to "be the source machines rely on, and be the obvious choice when the buyer is ready to act".

## How do answer engines decide what to cite?

Answer engines pick sources they can reach, understand and trust, and they prefer passages that directly support the point they are making. No engine publishes a full recipe, but the broad mechanics are well understood.

[](/posts/seo-in-the-ai-era/how-ai-answers-are-built.svg)

**They work from two kinds of knowledge.** The first is what the model learned during training, a blurred memory of a very large amount of text. This shapes which brands a model "knows" without looking anything up. The second is what it retrieves live, by searching and reading pages at the moment you ask. Most answers about businesses, prices and places lean on the second kind.

**Retrieval starts with search.** When an assistant looks something up, it runs one or more searches and reads the top results. Google has described breaking a question into several related searches and combining what comes back. This means a page that ranks for a sub-question can be cited even if it does not rank for the main one.

**Passages matter more than pages.** The model needs a short stretch of text that supports a specific claim. A page with one clear paragraph on the point is easier to use than a page that spreads the same idea over ten paragraphs of scene-setting.

**Agreement across sources builds confidence.** If your site, your business listing, a trade directory and three review sites all describe you the same way, a model can state it plainly. If they disagree about your address, your services or your prices, the model either hedges or leaves you out.

**Other people's words count.** Models learn about a brand largely from what others write: reviews, comparisons, news coverage, forum threads and "best of" lists. Our guide to [brand mentions in AI search](/blog/brand-mentions-ai-search) covers why a mention without a link still carries weight.

It also helps to know what does not exist. Google has said there is no special markup or file required to appear in AI Overviews, and that the normal fundamentals apply. There is no switch to flip. There is only the steady work of being readable, clear and well attested. For the mechanics of Google's own summaries, see [how Google AI Overviews work](/blog/google-ai-overviews).

## What is the playbook for winning both search and answer engines?

The playbook has seven steps, in this order: make the site readable, learn the questions, answer them first, add structured data, prove who you are, earn mentions elsewhere and measure both layers. Each step supports the next, so resist the urge to skip ahead.

[](/posts/seo-in-the-ai-era/seven-step-playbook.svg)

### Step 1: make your site readable by machines

Nothing else works if crawlers cannot reach and read your pages. Start here.

1. Check that important pages are indexed, using the URL Inspection tool in Google Search Console.

2. Make sure the main content is in the HTML, not hidden behind a script, a login or a click.

3. Review your robots.txt file. It controls which crawlers may visit, including named AI crawlers such as GPTBot, ClaudeBot and PerplexityBot. If you block them, you cannot be read by those tools.

4. Fix slow pages. The Core Web Vitals targets are a Largest Contentful Paint of 2.5 seconds or less, an Interaction to Next Paint of 200 milliseconds or less and a Cumulative Layout Shift of 0.1 or less.

5. Give every page one clear purpose and one address, with no duplicates competing against each other.

A [technical SEO audit](/blog/technical-seo-audit) covers the full list, and our guide to [AI crawlers and llms.txt](/blog/ai-crawlers-and-llms-txt) explains what each control does. One caution: llms.txt is a proposed convention, not a standard, and major AI providers have not confirmed that they use it. It is harmless to add, but robots.txt is what crawlers actually respect.

### Step 2: learn the questions your buyers ask

Keyword research still matters, but the unit of work is now the question, not the keyword. You need the short terms people type and the longer questions they ask an assistant.

Collect questions from four places:

- **Your own inbox and phone.** The questions customers ask before they buy are the questions they ask machines.

- **Search data.** Google Search Console shows the queries you already appear for. Tools such as Semrush and Ahrefs show what else people search around your topic.

- **The results page.** People Also Ask boxes show the follow-up questions Google associates with a topic.

- **AI tools themselves.** Ask an assistant the way a buyer would, and note the follow-up questions it suggests.

Group the questions by what the person wants to do: learn, compare, find a place or buy. Our guides to [keyword research](/blog/keyword-research) and [search intent](/blog/search-intent) go through the method.

### Step 3: answer first, then explain

Put the answer directly under the heading, in one or two sentences, and expand afterwards. This is the single most useful writing habit for both snippets and AI answers.

A good answer passage has four qualities:

- It sits under a heading phrased as the question.

- It answers in the first sentence, without a warm-up.

- It makes sense when lifted out of the page, so it names its subject instead of saying "it" or "this".

- It is followed by detail: steps, a table, an example or a caveat.

Compare two openings for a page about selling a house. The weak version begins, "Selling a home is one of the biggest decisions you will ever make." The strong version begins, "Selling a house has four stages: valuation, marketing, accepting an offer and conveyancing, and conveyancing usually takes the longest." The second one can be quoted. The first cannot.

This structure is also how you win [featured snippets](/blog/featured-snippets), which often feed voice answers and AI summaries.

### Step 4: add structured data

Schema markup is code that states plainly what a page is about: a business, a product, a person, an article, a set of questions. It removes guesswork for machines.

At a minimum, most businesses should mark up:

- **Organization or LocalBusiness** on the home or contact page, with name, address, phone, opening hours and links to official profiles.

- **Article or BlogPosting** on guides and posts, with author and dates.

- **Product and Offer** on anything with a price.

- **FAQPage** on pages with genuine questions and answers.

- **BreadcrumbList** so the site's structure is explicit.

Be realistic about what markup does. Google restricted FAQ rich results to well-known government and health sites in 2023, so FAQ markup rarely changes how your result looks. It still helps machines understand the page. Markup must also match what is visible on the page. Marking up reviews or prices that a visitor cannot see is against Google's guidelines. Our [schema markup guide](/blog/schema-markup) explains which types are worth the effort.

### Step 5: prove who you are

Machines are more willing to rely on a source when it is clear who is behind it and why they are qualified. Google's quality guidelines describe this as E-E-A-T: experience, expertise, authoritativeness and trust. It is not a score, but it is a useful checklist.

- Name the people behind the business, with real biographies and qualifications.

- Show first-hand experience: your own photos, your own numbers, your own process.

- Keep an accurate About page and a contact page with a real address and phone number.

- Use the same business name, address and phone number everywhere it appears online.

- Date your content and update it when facts change.

The last point matters more than it used to. Out-of-date facts on your own site are a common reason an AI tool describes a business wrongly. Our guide to [E-E-A-T](/blog/e-e-a-t) has the detail.

### Step 6: earn mentions on other sites

What others say about you is the strongest signal you do not control directly. Links still matter for ranking, and plain mentions matter for how models describe you.

Work on the sources your market already trusts:

- **Reviews.** Ask every satisfied customer, make it easy and reply to all of them.

- **Directories and listings.** Claim and complete the ones that matter in your field.

- **Local and trade press.** Offer data, commentary or a story worth covering.

- **Comparison and "best of" pages.** Find the lists that already rank for your category and ask what it takes to be considered.

- **Communities.** Take part in forums and groups where your buyers ask questions, as a helpful person and not as an advert.

Avoid shortcuts such as paid links and fake reviews. They put the business at risk and tend to be found out. See [link building](/blog/link-building) for honest methods.

### Step 7: measure both layers

You need the classic search numbers and a record of what AI tools say. One without the other gives you half the picture.

For search, track impressions, clicks, average position and conversions in Google Search Console and your analytics tool. For AI, write a fixed list of ten to thirty questions a buyer would ask, run them through the main assistants on a schedule and record three things: whether you are mentioned, how you are described and which sources are cited. The later section on measurement covers this in more detail.

## How can each industry capitalise on SEO in the AI race?

Every industry uses the same seven steps, but the questions, the trusted sources and the most useful schema types differ. The table gives the short version, and the sections below give the detail.

| Industry | What buyers ask machines | Where to focus first | Useful schema types |
| --- | --- | --- | --- |
| Real estate | Which area, which agent, what will it cost | Neighbourhood guides and agent profiles | RealEstateAgent, FAQPage |
| Small restaurants | Where to eat nearby for a specific need | Business profile, menu as text, reviews | Restaurant, Menu |
| Local trades | Who can fix this, how much, how soon | Service pages with prices and areas | Plumber, Electrician, HVACBusiness |
| Law firms | Do I have a case, what happens next, what are the fees | Plain-language guides by named solicitors | LegalService, Attorney |
| Clinics and dentists | Is this treatment right for me, what does it involve | Treatment pages reviewed by clinicians | Dentist, MedicalClinic |
| Online shops | Which product suits my need and budget | Complete product data and comparison content | Product, Offer, AggregateRating |
| Software companies | Which tool fits my team and what does it replace | Comparison pages, documentation, review sites | SoftwareApplication, Organization |

### Real estate

Estate agents rarely outrank the large property portals for individual listings, so the opening is local knowledge. A portal knows the number of bedrooms. You know which streets flood, which school catchments are changing and how long homes on a given road take to sell.

Buyers and sellers now ask assistants questions such as "is this area good for families", "what are the hidden costs of buying a flat" and "which agents near me are best for probate sales". Those questions reward specific, first-hand pages.

- **Write a guide for every neighbourhood you serve.** Cover transport, schools, typical property types, what tends to sell and who it suits. Use your own photos and observations.

- **Publish your process and your fees.** A clear page on how a sale works with you, with timescales and costs, is exactly the passage an AI answer needs.

- **Build a profile page for each agent.** Include areas covered, years of experience, specialisms and reviews. Mark it up with RealEstateAgent schema.

- **Answer the money questions.** Stamp duty, survey types, leasehold costs and chain delays are asked constantly.

- **Keep your Google Business Profile and portal profiles consistent.** The same name, address, phone number and service areas everywhere.

In the first month, pick your three strongest neighbourhoods and write the best guide available for each. That is a small, achievable start that no portal will match.

### Small restaurants

For a small restaurant, most of the battle happens off the website. When someone asks for "a quiet Italian place near the station with vegan options", the answer is assembled from your business profile, your menu, your reviews and local listings.

- **Complete your Google Business Profile.** Categories, opening hours, special hours for holidays, attributes such as outdoor seating or step-free access, photos and a booking link. Check it monthly.

- **Publish your menu as text on your own site.** A menu that exists only as a PDF or a photo is hard for machines to read. List dishes, prices and dietary notes in plain HTML, and mark the page up with Restaurant and Menu schema.

- **State the facts people filter by.** Dietary options, children's menu, dog policy, parking, private hire, average price per head. Each of these is a question someone asks an assistant.

- **Earn and answer reviews.** Reviews are where models learn what you are known for. A reply that mentions the dish and the occasion adds useful detail.

- **Get onto local lists.** Local bloggers, city guides and "best brunch" round-ups are heavily used as sources. Invite the writers in.

You do not need a blog. You need accurate facts in every place a machine might look, and a steady flow of real reviews. Our [local SEO guide](/blog/local-seo) covers the map results in detail.

### Local trades and home services

Plumbers, electricians, roofers and heating engineers are found in moments of need. The questions are urgent and practical: "who can fix a leaking boiler today", "how much does a consumer unit replacement cost", "is it worth repairing a ten year old boiler".

- **Create one page per service and one per area you genuinely cover.** Each page should say what the job involves, a typical price range, how long it takes and what affects the cost.

- **Publish prices where you can.** Trades that state a range are easier for an assistant to recommend than trades that say "call for a quote".

- **Show the work.** Before and after photos with a short description of the problem and the fix are strong first-hand evidence.

- **List your accreditations and insurance.** Name the scheme and your registration, so the claim can be checked.

- **Answer "repair or replace" questions.** These are high-intent questions that most competitors never write about.

- **Use the specific schema type.** Schema.org has types such as Plumber, Electrician and HVACBusiness, which are more precise than a general LocalBusiness.

Keep your listings on trade directories current as well, because assistants often lean on them when comparing local firms.

### Law firms and professional services

Legal, accounting and financial questions fall into the category Google treats with extra care, because bad advice can cause real harm. Trust signals carry more weight here than in almost any other field.

People ask assistants things like "what happens if I die without a will", "how long does an uncontested divorce take" and "do I need a solicitor to buy a house". They want a plain answer and a sense of what it will cost.

- **Write plain-language guides to the processes you handle.** Explain the stages, the timescales and the documents involved, without the jargon.

- **Put a named, qualified author on every page.** Link to a profile with their qualifications, regulator registration and areas of practice.

- **Be open about fees.** Fixed fees, typical ranges or at least how fees are calculated. Vague pricing pages are skipped.

- **Keep content current.** Laws and thresholds change. Date every guide and review it on a schedule, a habit covered in our guide to [content decay](/blog/content-decay).

- **State the limits.** A clear note on which jurisdiction the guide applies to helps a model use it correctly.

- **Build presence on legal directories and in local press.** These are the third-party sources models draw on for this field.

Use LegalService or Attorney schema for the firm and its people, and FAQPage markup on question pages.

### Clinics and dental practices

Health content is held to the highest standard of all. A practice should aim to be the clearest local source on the treatments it offers, written or reviewed by the clinicians who provide them.

- **Create a page for each treatment.** Cover who it suits, what happens during the appointment, recovery, risks, alternatives and price.

- **Show who reviewed the page.** Name the clinician, their qualifications and their registration, and add the review date.

- **Answer the anxious questions.** "Does it hurt", "how long will I be off work" and "what if it goes wrong" are the questions patients take to an assistant.

- **Avoid claims you cannot support.** Do not promise outcomes. State what the treatment usually involves and where results vary.

- **Keep practical facts exact.** Opening hours, emergency arrangements, accessibility, payment plans and whether you are taking new patients.

- **Ask for reviews in line with your regulator's rules.** Reviews shape how a practice is described.

Schema.org offers Dentist and MedicalClinic types, which suit this kind of site better than the generic business type.

### Online shops

In e-commerce, the assistant is turning into a shopping adviser. People ask "what is the best carry-on bag for a week away under a hundred pounds" and expect a short list with reasons. Your product has to be easy for a machine to understand and compare.

- **Complete your product data.** Dimensions, materials, weight, compatibility, care instructions, warranty and delivery terms. Missing attributes mean missing answers.

- **Mark up every product page.** Product, Offer and AggregateRating schema, with price and availability kept in step with the page.

- **Write for the use case, not only the item.** Buying guides such as "how to choose a carry-on bag" and comparisons between your own ranges answer the question behind the purchase.

- **Collect reviews with detail.** Reviews that mention how the product was used give models the language buyers use.

- **Keep your product feed accurate.** Google's shopping results rely on the feed you submit through Merchant Center.

- **Be findable on third-party lists.** Independent reviews and gift guides are common sources for product recommendations.

Category pages deserve attention too. A short, useful introduction that explains how the range differs helps both shoppers and machines.

### B2B software and services

Software buyers increasingly start with an assistant: "what are the alternatives to this tool for a team of twenty" or "which payroll software integrates with our accounting package". The answer is drawn from comparison pages, review sites, documentation and community discussion.

- **Publish honest comparison pages.** Explain who your product suits and who it does not. Balanced pages are more quotable than sales copy.

- **Make pricing and integrations public.** If a model cannot find your price or your integration list, it will describe a competitor that publishes theirs.

- **Keep documentation open and crawlable.** Good documentation answers the specific technical questions buyers ask late in a decision.

- **Maintain your profiles on review sites.** These are among the sources assistants lean on most for software.

- **Define your category in your own words.** A clear page on what your type of product is, and when to use it, builds [topical authority](/blog/topical-authority).

- **Check what models say about you.** Outdated descriptions of features and pricing are common and fixable.

Our guide to [getting cited in AI answers](/blog/how-to-get-cited-in-ai-answers) sets out the process step by step.

## What should you do in the first 90 days?

Spend the first month on foundations, the second on content and the third on reputation and measurement. A small team can make visible progress in that time if it keeps to the order.

[](/posts/seo-in-the-ai-era/first-90-days.svg)

**Days 1 to 30: fix the foundations.**

1. Run a technical audit and fix anything that blocks crawling or indexing.

2. Check robots.txt and decide which AI crawlers you allow.

3. Correct your business name, address, phone number and opening hours everywhere.

4. Complete your Google Business Profile if you serve a local area.

5. Ask the main AI assistants ten buyer questions and save the answers as your baseline.

**Days 31 to 60: answer the questions.**

1. List the thirty questions buyers ask most, grouped by intent.

2. Rewrite your most important service or product pages so each opens with a direct answer.

3. Publish five to ten new pages for questions you do not yet cover.

4. Add structured data to your key page types and test it.

5. Link related pages together, as set out in our guide to [internal linking](/blog/internal-linking).

**Days 61 to 90: build reputation and measure.**

1. Start a simple, repeatable way to ask customers for reviews.

2. Claim or update your profiles on the directories and review sites that matter in your field.

3. Pitch one useful story, data set or expert comment to local or trade press.

4. Repeat the ten AI questions from the first month and compare the answers.

5. Review search data and decide the next quarter's priorities.

The plan is deliberately plain. Most businesses that struggle in AI answers have not been outsmarted. They have gaps in the basics.

## What mistakes should you avoid?

The most common mistakes come from treating AI search as a trick to be gamed instead of an audience to be served.

- **Publishing large volumes of generic AI-written pages.** Content that says what every other page says gives machines no reason to choose it.

- **Blocking AI crawlers by accident.** A broad rule in robots.txt or a security setting can shut out the tools you want to be cited by.

- **Hiding facts in images and PDFs.** Menus, price lists and timetables should be text on the page.

- **Chasing a special file or tag.** There is no markup that guarantees inclusion in AI answers.

- **Letting facts drift.** Old prices, old addresses and discontinued services lead to wrong answers about you.

- **Ignoring third-party sites.** Your own site is only part of what a model reads.

- **Measuring only traffic.** Visits may fall while enquiries hold or grow, because buyers arrive better informed. Watch leads and sales as well.

## How do you measure progress?

Measure three things: visibility in search, presence in AI answers and business results. Report them together, because each explains the others.

| Layer | What to track | Where to find it |
| --- | --- | --- |
| Search visibility | Impressions, clicks, average position, indexed pages | Google Search Console |
| Local visibility | Profile views, calls, direction requests | Google Business Profile |
| AI presence | Mentions, how you are described, sources cited | A fixed prompt set, run on a schedule |
| Business results | Enquiries, bookings, sales, branded searches | Analytics, your booking or sales system |

For the AI layer, keep the method steady. Use the same questions, the same tools and the same schedule, and record the full answers with dates. AI answers vary from one run to the next, so look for patterns across several runs instead of reacting to a single response.

Pay attention to branded search as well. When more people search for your business by name, it often means they met you first in an answer and came looking. That is the AI era's version of word of mouth.

## Frequently asked questions

### Is SEO dead because of AI?

No. AI answer engines read the open web, and pages that are crawlable, clear and well regarded are the ones they draw on. What has lost value is thin content that repeats what everyone else says. Useful, specific pages matter as much as ever.

### What is the difference between SEO, AEO and GEO?

SEO aims to rank your pages in search results. AEO, or answer engine optimization, aims to have your content used as the direct answer in features such as snippets and AI Overviews. GEO, or generative engine optimization, aims to have your brand mentioned and cited by tools such as ChatGPT and Perplexity.

### How can a small business compete with big brands in AI answers?

A small business competes on specifics. Large brands tend to publish general content, while a local business can publish exact prices, real examples, local knowledge and first-hand experience. Assistants answering a specific question need specific sources, and that is where smaller firms have the advantage.

### Do I need special markup to appear in AI Overviews?

No. Google has said there is no special markup or file needed to appear in AI Overviews. A page needs to be indexed and eligible to show a snippet, and the usual SEO fundamentals apply. Structured data still helps machines understand what a page is about.

### Should I block AI crawlers from my site?

Usually not, if you want to be mentioned in AI answers. Blocking a crawler in robots.txt stops that tool from reading your pages, which makes it harder to be cited. Publishers who sell access to their content may decide differently, so treat it as a business decision.

### How long does it take to see results?

Technical fixes and business profile updates can show within weeks. New content and reputation work usually take several months to affect rankings. AI answers can change faster or slower, depending on how often each tool refreshes its sources, so track them over several months.

### Can I use AI to write my content?

You can use AI to help with drafts, outlines and editing, but the substance should come from you. Google's guidance focuses on whether content is helpful and reliable, not on how it was produced. Content with no first-hand detail or expert review is unlikely to earn rankings or citations.

### How do I check what AI says about my business?

Ask the main assistants the questions a buyer would ask, without naming your business, and note whether you appear. Then ask about your business directly and check the facts. Save the answers with dates, and repeat the same questions monthly to see the trend.

The businesses that do well in the AI era are not the ones with the cleverest tricks. They are the ones that are easy to read, clear about what they offer and well spoken of by others. If you would like help working out where you stand, [start a conversation](/contact).