Answer engine optimization (AEO) is the work of making your website easy for AI tools like ChatGPT, Perplexity and Google's AI Overviews to find, understand and cite when someone asks a question your business can answer. It starts with the same work as SEO, but the finish line moves: instead of a blue link on a results page, the goal is a sentence in an answer that names you.
Most of what you read about AEO is written for software companies with a marketing team and a budget for tracking tools. This guide is for everyone else: the dentist, the roofer, the accountant, the shop with one website and no one whose job is "search." Everything in it can be done without buying anything, and most of it in a few hours.
What is answer engine optimization?
An answer engine is any tool that responds to a question with a written answer instead of a list of links. Today that means Google's AI Overviews and AI Mode, ChatGPT, Perplexity, Gemini and Microsoft Copilot. When they answer, they often cite sources: a link, a business name, a short quote.
Answer engine optimization is everything you do to become one of those sources. Some of it is technical, like making sure these tools are allowed to read your site at all. Most of it is editorial: writing pages that answer real questions plainly enough that a machine can lift the answer out and trust it.
The name has competition. You will also see generative engine optimization (GEO), a term from a 2023 research paper by Pranjal Aggarwal and colleagues that measured which writing changes made pages more likely to appear in AI-generated answers (GEO: Generative Engine Optimization (opens in a new tab), presented at KDD 2024). You will see "LLM SEO" and "AI SEO" too. They describe the same goal. I use AEO because it names what changed: people now ask for answers, not pages.
How is AEO different from SEO?
They overlap more than the marketing suggests. An AI answer engine usually starts by searching the web, and it can only cite what it can find. So the basics of SEO, being crawlable, being indexed and having pages about the things people search for, are still the ground floor.
What changes is what "winning" looks like:
| Classic SEO | Answer engine optimization | |
|---|---|---|
| The prize | A high position in a list of links | A mention or citation inside a written answer |
| What gets judged | The whole page | Often a single passage or fact on the page |
| What earns it | Relevance, links and a good page | Clear, specific, quotable answers, plus being mentioned consistently elsewhere |
| How you measure it | Rankings, clicks, Search Console | Checking the answers themselves, AI referral traffic, branded search |
The biggest practical difference is the unit of competition. Google ranks pages. An answer engine quotes passages. A page can rank tenth and still be the one quoted, if it holds the clearest sentence on the exact point being asked.
How do AI answer engines choose what to cite?
Each tool works a little differently, and none publishes a full recipe. But the broad pattern is shared, and it is enough to act on:
- They search. For most questions about the real world, the assistant runs searches first. Google's AI features draw on Google's own index. ChatGPT search draws on third-party search providers, including Microsoft Bing, along with OpenAI's own crawler, called OAI-SearchBot (Searching the web with ChatGPT (opens in a new tab)).
- They read a handful of pages. The results are fetched and read, often quickly and often without running the page's JavaScript.
- They pull out passages. The answer is assembled from the parts of those pages that address the question most directly.
- They cite what they used. Sources that supplied a specific fact or a clear explanation get the link.
Two conditions follow from that. You have to be readable by the tool at step 2, and you have to hold a clear, specific answer at step 3. Most small-business sites fail one or the other without knowing it.
It also means your visibility in Bing matters more than it used to. If you have never checked whether Bing has your site, Bing Webmaster Tools (opens in a new tab) is free, and it can import your setup from Google Search Console in a few minutes.
What can a small business do about it?
Here is the checklist I work through on client sites, in the order I check it: the first items can hide every later one. On a typical small-business site, the whole list is a few hours of work plus some writing, spread over a few weeks.
1. Make sure AI tools are allowed to read your site
This is the one people skip, and it is the one that silently zeroes out everything else. Check three things:
- Your robots.txt file (at yoursite.com/robots.txt). It should not block Googlebot, Bingbot, OAI-SearchBot or PerplexityBot. Blocking GPTBot, which collects training data, is a reasonable choice and does not stop ChatGPT's search from finding you.
- Your firewall or CDN settings. Services that sit in front of a website, Cloudflare being the most common, now offer switches that block AI crawlers. In July 2025 Cloudflare announced it was changing the default to block AI crawlers (opens in a new tab). Look for a bot or AI crawler setting in your dashboard and make sure the search crawlers you want can get through.
- Your security plugins. On WordPress, some security plugins block unfamiliar crawlers by default. If a tool can't load your pages, it can't cite them.
2. Put the important facts in plain text
In December 2024, Vercel analyzed traffic from AI crawlers and found that none of the major AI crawlers run JavaScript (opens in a new tab), including those from OpenAI, Anthropic and Perplexity. Google's is the exception. If your prices, services, hours or service area only appear after a script runs, or live inside an image or a PDF, many assistants simply don't see them.
The fix is boring and effective: the facts a customer would ask about should be written out as text on a normal page. If your site is built in a way that renders content in the browser, ask whoever built it whether the important pages are rendered on the server. Mine are built as plain HTML for exactly this reason.
3. Answer the question in the first two sentences
An answer engine is looking for the passage that answers the question. Make it easy to find:
- Open each page, and each section, with the answer. Explain afterward.
- Use headings that match how people ask ("How much does a new roof cost?"), not clever labels ("Investing in your home").
- Keep one topic per section, so a quoted passage makes sense on its own.
- Be specific. "Most repairs cost $300 to $900, depending on access and parts" is quotable. "Pricing varies" is not.
This is also simply better writing for people, which is why it holds up whatever the tools do next.
4. Write the pages your customers' questions deserve
Most small-business sites have a home page, an about page and one page listing every service. Answer engines, like search engines, need a page that is about the specific thing being asked. If people ask "do I need a permit for a fence in my town," the business with a clear page answering that question is the one that gets named.
A good starting list is the questions you answer on the phone every week. Each one can be a section of a service page, an entry in an FAQ, or its own article.
5. Be described the same way everywhere
Assistants cross-check. When your business is described consistently in several places, such as your site, your Google Business Profile, industry directories, local news or a partner's website, you become a safer source to name. When your name, address, phone number or list of services differs from one place to the next, you become a riskier one.
For local businesses, this means a complete Google Business Profile, a steady flow of genuine reviews, and accurate listings on the directories that matter for your trade. It is the same work that helps you in the map results, which is a good reason to do it first.
6. Use structured data to remove doubt, not to chase a trick
Structured data (schema.org markup) describes your business in a format machines read directly: your name, address, service area, hours, what you sell. Google is clear that no special markup is required for its AI features, so don't believe anyone who sells schema as a secret AEO switch. Its value is that it removes ambiguity about who you are, which helps every system that reads your site, AI or not.
7. Consider an llms.txt file
An llms.txt (opens in a new tab) file is a short, plain-text map of your site written for AI tools, placed at yoursite.com/llms.txt. Jeremy Howard proposed it in September 2024. It is not a Google ranking factor, and Google says no such file is needed for its AI features. It is cheap to add, though, and it helps AI agents and developer tools that read it find your most useful pages. This site has one, if you want to see what it looks like.
Who can skip this for now?
Not every business needs to spend time on AEO yet. You can leave it for later if:
- Nearly all your work comes from referrals, and new customers look you up by name rather than asking for a recommendation.
- Your site is still missing the basics: a clear list of services, prices or price ranges, a service area and a way to get in touch. Fix those first; they are also most of what an answer engine needs.
- You sell to a handful of large customers you already know by name.
If customers find businesses like yours by asking a question, like "who fixes ___ near me" or "what should I look for in a ___," AEO is worth the few hours.
What nobody knows yet
None of these tools publishes how it chooses sources, and they change often. What I can say with confidence is what each company documents about its own crawlers and eligibility, and what I have seen hold up on the sites I manage. What nobody can tell you yet is how much weight any single factor carries, or how long a change takes to show up in answers. Be wary of anyone who claims to know. This guide sticks to steps that are documented, cheap and useful for human readers too, so none of it is wasted if the tools change.
What about AEO tools?
There is a growing market of tools that track how often your brand appears in AI answers. Ahrefs, Semrush, HubSpot and specialists like Profound all sell some version of it. They are useful for measuring at scale, and I use one for my own tracking.
They are trackers, not fixes. None of them makes you more quotable; they tell you whether the work above is paying off. If you are a small business, start with the checklist and the free checks below, and only pay for tracking once you have something to track.
How do you know if it's working?
You can measure the basics for free, once a month, in about fifteen minutes:
- Ask the assistants yourself. Write down five questions a customer would really ask, like "who does ___ near Jackson, MS" or "how much does ___ cost," and put them to ChatGPT, Perplexity and Google. Note who gets named and which pages are cited. Use the same questions each month so the results are comparable.
- Check your analytics for AI referrals. Visits from chatgpt.com, perplexity.ai and gemini.google.com show up as referral traffic. They are usually small numbers, and they undercount: apps and some browsers strip the referrer, so part of that traffic lands in your "direct" visits instead. Treat what you see as the floor, not the total.
- Watch branded search in Search Console. When assistants mention you, some people search for your name afterward. A rise in searches for your business name is often the first sign.
Expect this to move slowly. Assistants refresh what they know on their own schedules, and answers vary from one run to the next. Judge the trend over months, not a single check.
Where I'm starting from
I am applying all of this to my own site in public, so here is my honest starting point. When I began tracking matchlessweb.com in Ahrefs in September 2026, it ranked for zero organic keywords. That was deliberate: the old version of my site ranked well for the audience I used to serve, and when I repositioned the business I rebuilt from scratch and let those pages and their rankings go, knowing exactly what it would cost. So this is a genuine cold start. I now track 20 questions a buyer might ask about web design, care plans, funnels and SEO, in Jackson and nationally, across ChatGPT, Google's AI Overviews and AI Mode, Gemini and Perplexity, and I check them every week.
This blog is part of that work. I will update this guide as the numbers move, and the date at the top will tell you when I last did.
The short version
- AEO is SEO with a different finish line: a mention inside an answer instead of a link in a list.
- Let the tools read you: check robots.txt, your firewall's AI crawler settings and any security plugins.
- Put important facts in plain text, and answer each question in its first two sentences.
- Build pages around the questions customers actually ask.
- Be described consistently across the web, starting with your Google Business Profile.
- Treat structured data and llms.txt as clarity, not tricks.
- Measure monthly by asking the assistants yourself, and judge the trend over months.