What is the difference between AEO and GEO?

AEO, answer engine optimisation, is about making content easy for systems that give direct answers to find and cite. GEO, generative engine optimisation, a term from a 2023 research paper, focuses on visibility inside answers written by generative AI search engines. In practice they overlap almost entirely.

Should I block AI crawlers in robots.txt?

Decide separately for training and for search. OpenAI and Anthropic both document separate crawlers for each, so you can decline model training and still allow the crawlers that power their search features. Check each vendor's own documentation for the current names.

How can a local business get mentioned in AI answers?

Complete your Google Business Profile and other map listings, and keep your name, address and phone number identical everywhere. Collect genuine reviews, create real service-area pages and add LocalBusiness structured data that matches the page.

What should a SaaS company publish to be cited by AI assistants?

Clear pages on what the product is, who it is for, pricing, alternatives and integrations, with real prices in text. Keep documentation public and crawlable, publish a dated changelog and write comparison pages that are honest about where competitors fit better.

How do app makers improve how AI assistants describe their apps?

Complete every app store listing with a plain-language description and give each app its own page on your website. Add public FAQs and support pages, and use the same name and description everywhere so the facts agree.

What helps publishers and content websites get cited?

Cover topics in depth through linked pages, make authors and the organisation behind the site clear, and cite primary sources. Publish original data only when you genuinely have it, and show published and updated dates.

How can an online shop appear in AI shopping answers?

Add Product structured data, put detailed specifications and honest descriptions in text, and keep returns and shipping pages clear. Keep merchant feeds in step with your site's prices and stock, and write category buying guides.

How do I know whether AI assistants mention my business?

Ask the main assistants the questions your customers would ask, and log the answers and cited sources each month. Also watch your analytics for referral visits from AI tools, bearing in mind that some visits arrive without a referrer.

How to get found in AI answers: a practical guide to GEO and AEO

To get found in AI answers, make yourself the easiest correct source to use. Let AI search crawlers reach your pages, answer questions clearly at the top of each page, describe your business with structured data, keep your facts identical everywhere they appear, and get mentioned by other credible sites. Nobody outside the AI companies knows exactly how their systems choose which sources to use, so anyone promising guaranteed placement is guessing. What follows is the sensible groundwork, then a checklist for each of five kinds of business.

What AEO and GEO mean

AEO (answer engine optimisation) means making your content easy for systems that give a direct answer, rather than a list of links, to find, understand and cite. That covers AI assistants, voice assistants and the answer boxes in search results.

GEO (generative engine optimisation) means improving how visible your content is inside answers written by generative AI search engines. The term comes from the research paper GEO: Generative Engine Optimization by Aggarwal et al. (2023).

In practice the two overlap almost entirely, and both sit on top of ordinary SEO. An assistant that searches the web before answering still has to find, fetch and trust your page. If the basics are not in place yet, start with our guide on how to get traffic from SEO.

What nobody knows

The systems behind AI answers are not public, differ between products and change often. Some answers come from what a model learned in training, some from a live web search with cited sources, and many from both. So there is no checklist that guarantees a mention. The realistic aim is narrower: when a system goes looking for an answer in your area, your page should be easy to reach, easy to read and clearly correct.

The core principles

1. Be the clearest source

Open each page with a paragraph that answers its main question directly, then give the detail.
Add a short key takeaways list near the top of longer pages.
Include a FAQ with the questions customers really ask, answered in two or three plain sentences.
Define your terms in one sentence the first time you use them.
State specific facts (prices, limits, supported platforms, delivery areas) in text, not only in images or downloadable PDFs.

2. Let AI search crawlers in, and decide about training separately

Your robots.txt file can allow or block crawlers by name. The main AI companies publish the names they use, and several now separate crawling for search from crawling for model training:

OpenAI: OAI-SearchBot is used for search results in ChatGPT, and GPTBot for training its models. OpenAI's crawler documentation says each setting is independent, so you can allow one and block the other.
Anthropic: ClaudeBot collects content that may be used for training, Claude-SearchBot indexes content to improve search results, and Claude-User fetches pages when a person asks Claude a question. See Anthropic's article on its crawlers.
Perplexity: PerplexityBot surfaces and links websites in Perplexity's search results, and Perplexity's bot documentation says it is not used to crawl content for AI foundation models.
Google: Google-Extended is not a separate crawler but a robots.txt control token. Google's crawler documentation says it governs whether content Google crawls may be used to train future Gemini models and for grounding in Gemini apps, and that it does not affect inclusion in Google Search.

The point to take away: blocking a training crawler is a separate choice from blocking a search crawler. You might reasonably decline training and still want to appear in AI search results, or the reverse. Make both choices deliberately, write them into robots.txt with a User-agent: line per crawler, and check your hosting or CDN settings too, because some offer switches that block AI crawlers regardless of robots.txt. Crawler names change, so check the vendor pages above rather than copying a list from a blog post, this one included.

3. Describe yourself with structured data

Structured data in the schema.org vocabulary states plainly what a page is. The types most businesses need are Organization, LocalBusiness, Product, SoftwareApplication and FAQPage. Mark up only what is visible on the page, and keep it in step with the page when things change. One note on FAQPage: Google's Search Central changelog records that FAQ rich results stopped appearing in Google Search in 2026, so do not expect a special snippet from it. We still use it because it describes the questions a page answers in a tidy, machine-readable form at no real cost.

4. Consider an llms.txt file

llms.txt is a proposal for a Markdown file at the root of a site that summarises what the site is and links to its most useful pages, so that AI agents can find the important material without wading through navigation and scripts. It is a proposed standard, not something the major AI engines have confirmed they use. It is cheap to write and doubles as a useful index of your own site, which is why we publish one.

5. Keep your facts consistent across the web

Use the same business name, description, prices, product names, address and contact details on your website, your listings, marketplace profiles, social profiles and partner pages. When sources disagree, an assistant has to pick one or hedge, and an old price on a forgotten directory can be repeated back to customers as fact. Keep a list of everywhere you are listed and check it when anything changes.

6. Get mentioned on other credible sites

Being described accurately by others, in reviews, partner and integration directories, trade publications and community discussions, gives assistants more than your own word to go on. As with links for SEO, these mentions have to be earned. Paid placements dressed up as independent reviews are a poor bet and damage trust when spotted.

7. Keep pages current, with visible dates

Show a published date and, where relevant, a last-updated date on each page, and use the matching datePublished and dateModified fields in your structured data. Change the date only when the content really changes. A page that is plainly current is easier to rely on than one that might be out of date.

What we do on mcp-g.com

We follow the same pattern here. Every blog post opens with an answer-first paragraph, carries key takeaways and a per-post FAQ, includes FAQPage and BlogPosting structured data, and has an entry in our /llms.txt file. We cannot tell you exactly how much each piece helps, and neither can anyone else, but each one makes the page easier to use correctly.

Local businesses

Claim and complete your Google Business Profile, plus the equivalent listings on Bing and Apple Maps and any directories your customers actually use. Follow Google's guidelines for representing your business.
Use exactly the same name, address and phone number everywhere, matching your website. Keep opening hours, including holiday hours, up to date.
Ask real customers for reviews and reply to them, the critical ones included. Never buy reviews or offer anything in exchange for good ones.
Create a page for each genuine service and each area you really serve, with real local detail. Do not copy one page and swap the town name; that is thin content and reads like it.
Add LocalBusiness structured data with your name, address, phone, opening hours and service area, matching what is on the page.
Answer the questions people ask before they call: price ranges, areas covered, lead times and what is included.

SaaS

Have one clear page for each of: what the product is, who it is for, pricing, alternatives and integrations.
Put real prices on the pricing page, in text. "Contact us" pricing leaves an assistant nothing to quote.
Keep documentation public and crawlable, in HTML rather than only in PDFs or behind a login.
Write honest comparison pages that say where a competitor is the better fit, date them, and correct them when either product changes.
Publish a dated changelog so the current feature set is easy to establish.
Use SoftwareApplication and Organization markup, with an offer for each price.
Keep your listings on integration marketplaces and review sites consistent with your own site.

Apps

Complete every app store listing: a plain-language description, accurate screenshots, the right category and current release notes. Use the same name and description you use elsewhere.
Give each app its own page on your website, the one place you fully control, saying what it does, who it is for, which platforms it works with, what it costs and how to install it. Our apps page follows this pattern.
Publish FAQs and support pages publicly: set-up, troubleshooting, data handling, cancelling and uninstalling.
Add SoftwareApplication markup with the operating system or platform, category and price.
Link the web page and the store listing to each other so the two are clearly the same product.

Websites and publishers

Build topical depth: cover a subject thoroughly through a set of linked pages rather than publishing one-off pieces on everything.
Make authorship clear, with bylines, author pages that explain relevant experience, an about page and contact details, plus matching Organization and Person markup.
Cite your sources and link to the primary material, so a reader or a machine can check the claim.
Publish original data only when you genuinely have it, such as your own records or measurements, and explain how it was gathered. Do not dress estimates up as research.
Make sure the main text is in the HTML your server sends, not only drawn in by scripts, so any crawler can read it.
Show published and updated dates on every article.

E-commerce

Add Product structured data with name, brand, identifiers such as GTIN where you have them, price, currency and availability.
Put detailed specifications in text: dimensions, materials, compatibility, what is in the box, care instructions.
Write honest descriptions that include limitations and who the product is not right for. Assistants answering "which should I buy" questions need that detail.
Keep clear, specific returns, shipping and delivery pages, and link to them from every product page.
Keep merchant feeds, such as Google Merchant Center and marketplace feeds, in step with your site's prices and stock.
Write category buying guides that answer "which one should I choose for this job" for each main range you sell.

How to measure it

There is no single report for AI visibility yet, so measure it two ways.

Ask the assistants directly. Once a month, ask the main assistants the questions your customers would ask: what your business does, the best options in your category, how much your product costs. Log the date, the assistant, the answer and any sources it cites in a spreadsheet. Answers vary from one run to the next, so look for patterns over months rather than reacting to a single reply. When an answer is wrong, find the source of the wrong fact and correct it there.
Watch referral traffic. In your analytics, look for visits referred from AI tools such as chatgpt.com, perplexity.ai, claude.ai and gemini.google.com. Some visits from AI tools may arrive without a referrer at all, so treat these numbers as a floor, not a total.

Search Console and Bing Webmaster Tools still matter too, because many assistants lean on web search and the same crawling and indexing problems will hold you back in both.

For the groundwork underneath all of this, read how to get traffic from SEO, or browse the rest of the blog.