Answer engine optimisation (AEO) is the work of getting a company named, recommended and described accurately when people ask AI assistants a question. When a buyer asks ChatGPT, Claude, Perplexity, Gemini or Google's AI Overviews which firm to hire or which product to buy, the answer usually names two or three options. AEO is how you become one of them.
This guide explains how those answers are put together, what you can influence, and how we approach the work at Cited. It is written for founders and marketing leads who want a clear picture rather than a list of tricks.
Why AI answers matter commercially
A search results page gives a buyer ten links and leaves the comparison to them. An AI assistant does the comparison itself and hands back a shortlist with reasons. If you are on that shortlist, you arrive in the buyer's mind already recommended. If you are not, you are often never considered, and nothing in your analytics tells you it happened.
That is the main difference worth holding on to. In classic search, a lower position still meant some visibility. In an AI answer, the firms that are not named are simply absent.
How an AI assistant builds an answer
The details differ by product, but most AI assistants follow the same broad path when a question needs current information:
- The question is broken down. The model turns one question into several searches. Google describes this "query fan-out" technique in its documentation for AI Mode, and the same pattern is visible in ChatGPT and Perplexity. A question about the best accountant for a startup in Leeds might become searches about startup accountants, Leeds firms, pricing and reviews.
- Pages are retrieved. Each search returns pages from a search index or the assistant's own crawler. Pages the crawler cannot reach or read never enter the process.
- Sources are weighed. The model favours pages that answer the question directly and claims that several independent sources agree on.
- The answer is written. The model names the options it is most confident about and often cites the sources it used.
Alongside this live retrieval, each model also carries what it learned in training. A company that has been described consistently across the web for years is already "known" before any search happens. A newer or less documented company depends almost entirely on what retrieval finds.
Which AI assistants matter
The main ones for most businesses are ChatGPT, Google's AI Overviews and AI Mode, Gemini, Perplexity, Claude and Microsoft Copilot. They do not read the same sources. Google's AI features draw on Google's own index, so strength in ordinary Google search tends to carry over. Perplexity shows its sources openly and leans on fresh pages and community discussion. ChatGPT runs its own search with its own crawler. The practical consequence is that a company can lead in one assistant and be missing from another, which is why we always measure across several.
The three layers of AEO work
We group the work into three layers. Each depends on the one beneath it.
1. Foundation: a site AI can reach, read and understand
AI crawlers need permission to visit your site and need to be able to read it once they arrive. In practice that means:
- Crawl access. Your robots.txt and any firewall or bot protection should allow the crawlers you want, such as OAI-SearchBot for ChatGPT search, PerplexityBot, Claude-SearchBot and Googlebot. Many sites block these by accident through a default security setting.
- Server-rendered content. Several AI crawlers do not run JavaScript. An analysis published by Vercel in December 2024 found that the crawlers behind ChatGPT and Claude fetched JavaScript files but did not execute them. If your key content only appears after scripts run, those crawlers see an empty page.
- Clear entity information. Who you are, what you do, where you operate and who runs the company should be stated plainly in text and in structured data, and should match what other sites say about you.
- Speed and stability. Pages that time out or return errors are skipped.
2. Answers: pages that answer the questions buyers ask
Most company websites describe the company. Buyers ask questions. The gap between the two is where most firms lose. Good answer pages:
- Address one real question per page, in the words buyers use.
- Give the direct answer early, then the detail.
- State specifics an AI assistant can repeat: who the service is for, where you operate, how the work runs, what it costs or what drives the cost.
- Show evidence: named people, case studies, credentials and dates.
Comparison pages, pricing explanations and "how to choose" guides are particularly useful because they match the questions buyers ask just before they decide.
3. Distribution: the third-party sources AI engines consult
AI assistants rarely rely on a company's own site alone. They look for agreement from independent sources: industry lists, directories, review sites, press coverage, forums and comparison articles. In our client work, the sources cited for a category are often a short, predictable list. Being accurately present on that list, with facts that match your own site, is frequently what moves a company from occasionally mentioned to consistently recommended.
What does not work
- Prompt stuffing and hidden text. Instructions aimed at AI models hidden in pages are easy to detect and can damage trust in the whole site.
- Mass-produced pages. Hundreds of thin pages targeting every phrasing of a question add noise, not authority.
- Guarantees. No one controls what an AI model says. Anyone promising a fixed position in ChatGPT is promising something they cannot deliver.
- Optimising for one assistant. Checking only the assistant you personally use gives you a sample of one.
How to measure progress
Start with a fixed set of the questions your buyers ask, run them across the main assistants, and record who is named, how each firm is described and which sources are cited. Repeat the same set on a regular schedule. Our guide on measuring AEO covers the metrics in detail.
Where to start this week
- Write down the ten questions a buyer would ask before choosing a firm like yours.
- Ask each one in ChatGPT, Gemini, Perplexity and Claude. Note who appears and which sources are cited.
- Check your robots.txt and bot protection settings for AI crawlers.
- Look at your most important pages with JavaScript turned off.
- Compare what your site says about you with your LinkedIn page, directory listings and review profiles.
Frequently Asked Questions
What is answer engine optimisation?
Answer engine optimisation (AEO) is the practice of getting a company named, recommended and described accurately in the answers AI assistants such as ChatGPT, Claude, Perplexity, Gemini and Google's AI Overviews give to buyers' questions.
Is AEO the same as GEO?
Mostly, yes. Generative engine optimisation (GEO), AI search optimisation and LLM optimisation are other names for the same goal. The terms are used interchangeably across the industry.
Does AEO replace SEO?
No. Much of the foundation is shared, and Google's AI features draw on Google's search index. AEO adds work on answer-shaped content, third-party sources and measurement across several AI assistants.
How long does AEO take to show results?
Technical fixes can be picked up within weeks once crawlers revisit the site. Changes driven by new content and third-party sources usually take a few months, and vary by category and by assistant.