Answer engine optimization is the practice of making your site easy for an AI answer engine to read, quote, and credit. A search engine sends people a list of links. An answer engine writes them a paragraph and cites a handful of sources. AEO is the work of being one of those sources.

No. Most AI answer engines build their answers on top of a search index, so a page that ranks well is a page that is easy to find and quote. Classic search still sends real traffic. AEO is an addition to SEO, not a replacement for it.

How long does AEO take?

Longer than a page edit, shorter than a link-building campaign. Crawlers have to re-fetch the page, and the answer engine has to decide the new version is worth quoting. Expect weeks, not days, and expect it to vary by engine. The honest answer is that nobody can promise you a timeline, and you should be wary of anyone who does.

Do I need a consultant to do AEO?

For most small sites, no. The work is structural: answer the question in the first sentence, use plain headings, put comparisons in a table, keep pages crawlable, and add basic schema. All of that is a content and markup job you can do yourself. Hire help when the bottleneck is writing volume or a genuinely complex site, not because the acronym sounds technical.

How do I measure AEO?

Two signals. First, which AI crawlers fetch which of your pages, which you can only see server-side because AI crawlers do not run the JavaScript that most analytics tools depend on. Second, whether people arrive on your site from an AI assistant afterwards. Kymo Field Check gives you a free one-time baseline from public signals, and Kymo tracks both signals continuously once installed.

The Small Business Guide to AEO (Answer Engine Optimization)

Answer engine optimization is the practice of making your website easy for an AI assistant to read, quote, and credit by name. When someone asks ChatGPT, Claude, Perplexity, or Google's AI results a question your business could answer, AEO is the work that decides whether your site is one of the sources it pulls from.

TL;DR

AEO means writing and structuring your pages so an AI answer engine can lift a clear answer out of them and cite you as the source.
It is not a replacement for SEO. Most answer engines sit on top of a search index, so ranking well and being quotable are the same job done twice.
The mechanics are unglamorous: be crawlable, answer the question in the first sentence, use plain headings, put comparisons in real tables, add basic schema, and keep pages current.
AI crawlers do not run JavaScript. Anything that only appears after a script runs is, as far as they are concerned, not on your page.
You cannot see any of this in Google Analytics. Measuring AEO needs a server-side view of which AI crawlers fetched which pages.

Why this matters now

A growing share of the questions that used to become a Google search now become a question typed at an assistant. The person asking gets a written answer with three or four citations underneath. They read the answer. Sometimes they click a citation. Often they do not.

That changes the shape of the funnel. Ten blue links gave ten businesses a chance at the click. A written answer gives roughly three sources a chance at the credit, and gives everyone else nothing at all. Being on page one of a search result and being quoted in an answer are different outcomes, and the second one is getting more valuable every quarter.

If you have been doing fine on Google for years, this is the part of your funnel you currently cannot see. Your rankings can hold steady while the number of people who ever run the search quietly falls. Your analytics will show you the traffic that arrived. It will not show you the answer that was written about your category last Tuesday, or whether your competitor was the one cited in it.

None of this means search is over. People still search, still click, and still buy from links. What changed is that a second discovery channel opened next to the first one, and most small businesses have no instrument pointed at it.

SEO vs AEO vs GEO

Three acronyms are in circulation for roughly two ideas. AEO stands for answer engine optimization. GEO stands for generative engine optimization. In practice they overlap so heavily that arguing about which one to use is a waste of a morning. The practice is the same: make the page answerable and quotable. SEO is the genuinely separate discipline, and it still works.

How the three compare
	SEO	AEO	GEO
Optimizes for	Ranking in a list of links	Being quoted in a written answer	Being quoted in a written answer
Who reads the result	A person scanning results	A model assembling an answer, then a person	A model assembling an answer, then a person
What winning looks like	Position 1 to 3 for a query	Named as a source under the answer	Named as a source under the answer
How you measure it	Rank trackers, Search Console, organic sessions	AI crawler hits per page, plus referrals from assistants	Same signals, different vocabulary
Still worth doing	Yes	Yes	Yes, and it is the same work as AEO

Treat AEO and GEO as one practice with two names. Spend the time you would have spent picking a side on the checklist further down.

How AI engines pick sources

An answer engine has to do three things before it can cite you: fetch your page, understand it well enough to extract a claim, and decide that claim is worth attaching your name to. Each step has a failure mode you control.

It has to be able to fetch the page

AI crawlers request your pages like any other bot, and your server can refuse them. Arobots.txt rule, a firewall setting, or a managed bot-protection default can block GPTBot or ClaudeBot without anyone at your company deciding to. Plenty of sites are invisible to answer engines for exactly this reason and have no idea. Check what your server actually returns to each crawler before you change anything else.

It has to see your content without running JavaScript

AI crawlers fetch HTML. They do not behave like a browser with a full rendering engine waiting for your framework to hydrate. If your page arrives as an empty shell and fills in from an API call, the crawler sees the shell. View the page source, not the inspector, and confirm the words you care about are in the HTML that comes back.

Answer-first content is easier to quote

A model extracting an answer wants a sentence it can lift whole. Give it one. Put the direct answer in the first sentence under the heading, then explain, then qualify. Writing that builds to a conclusion in paragraph six reads fine to a person and gives a model nothing clean to pull.

Structure carries meaning

Headings that state the question being answered, short paragraphs, real bulleted lists, and one h1 per page all make the document easier to parse. A comparison written as three paragraphs of prose is harder to extract than the same comparison in a table with a header row. Use the table.

Schema helps, and it is cheap

Structured data in JSON-LD tells a machine what kind of thing your page is. Article, FAQPage, Product, LocalBusiness, and BreadcrumbList cover most small business needs. The vocabulary is documented at schema.org and it takes an afternoon to add. Schema does not make a weak page strong. It removes ambiguity from a good one.

Be the source, not the summary

Models cite pages that contain something the rest of the internet does not already say. Your prices, your process, your data, your specific answer to the question your customers actually ask on the phone. A page that restates the top three results for a query is competing with those three results on their terms and losing.

Freshness is a signal

A dated page that clearly says when it was last updated, and was actually updated, is easier to trust than one that could be from any year. Put a visible updated date on pages that make factual claims, and keep it honest.

Getting started: seven steps for this week

Work through these in order. Each one names what to do and how to confirm you did it. None of them require a developer for a typical small business site, though step two is easier with one.

Check that AI crawlers can reach you. Open your robots.txt and look for rules naming GPTBot, ClaudeBot, PerplexityBot, Google-Extended, or a blanket disallow. Verify: load yourdomain.com/robots.txt in a browser and read it. If a bot you want is disallowed and you did not intend that, remove the rule. If you use Cloudflare or a similar service, check its bot settings too, because several of them now block AI crawlers by default.
Confirm your pages render without JavaScript. Verify: right click, view page source, and search that raw HTML for a sentence from the middle of the page. If it is missing, your content is being assembled in the browser and crawlers are seeing an empty page. Server-side rendering or static generation fixes it.
Rewrite the first sentence of your five most important pages. State the answer plainly before anything else. Verify: read only the first sentence aloud. If it does not answer the heading above it, it is not done.
Turn one comparison on your site into a table. Pricing tiers, plan differences, your service versus the obvious alternative. Give it a real header row. Verify: the table has a thead and each cell reads as a full answer on its own, so "Yes" and "No" are written out rather than implied by a tick graphic.
Add schema to your key pages. Article or FAQPage for content, LocalBusiness for a physical business, BreadcrumbList everywhere. Verify: run each page through Google's Rich Results Test and confirm it detects the type you added with no errors.
Write the page that answers the question you get asked most. The one you answer on the phone twice a week. Title it as the question. Answer it in the first sentence. Verify: the page exists at a clean URL, is linked from somewhere on your site, and appears in your sitemap.
Measure it. Get a baseline now, before the changes above have had time to take effect, so you can tell later whether anything moved.

Measuring AEO

What does not get measured does not get improved, and AEO is unusually easy to leave unmeasured because the standard tools are blind to it.

Google Analytics 4 cannot show you AI citations. It is a client-side tool: it runs a script in a visitor's browser and reports what that browser did. AI crawlers do not run that script, so they leave no trace in GA4 at all. A page can be fetched hundreds of times by an answer engine and appear in GA4 as zero. Referrals from AI assistants sometimes show up, though attribution is inconsistent and a referral only happens after someone clicks, which is the minority case.

Two signals are worth watching, and both need a server-side view:

Which AI crawlers fetch which of your pages. This is the earliest signal you get. It tells you an answer engine considered your page relevant enough to read, weeks before any of it turns into a visit.
Whether anyone arrives from an AI assistant afterwards. This is the signal that connects to revenue.

Kymo Field Check is a free way to get the first read. Give it a URL and it audits that site from public signals only, so the site does not need anything installed. It checks what your server actually returns to each of the major AI crawlers, whether your robots.txt allows or blocks them, whether your content survives without JavaScript, and what machine-readable structure you have. The report arrives by email. No account required.

Kymo itself is the continuous version. It is a cookieless analytics tool that tracks human visitors and AI crawlers in one dashboard: which pages assistants fetch, the likely prompts behind those fetches, and confirmed click-throughs when an AI referral turns into a real visit. Solo is $9 a month for up to 10,000 events, Studio is $29 a month for up to 100,000, every feature is on both plans, and AI crawler tracking does not count against your event limit. Both plans start with 14 days free and no card.

Kymo measures and reports. It does not write your pages or promise you a citation, because nobody can promise that. What it can do is tell you whether the work in the checklist above changed anything.

Get your baseline

Run a free Field Check on your own site. One URL, no account, no snippet to install. You get a report on what each AI crawler sees when it visits you.

Run a free field check See the live demo

Already measuring, and the numbers are flat

If your AI Visibility report shows few fetches, or fetches that never turn into click-throughs, the problem is usually one of three things, in this order of how often it turns out to be the cause.

Nothing is being crawled at all. Check step one above. A blocked crawler produces exactly this picture, and it is the fastest thing to rule out. Run a Field Check and read the AI crawler access section, which tells you what your server returns to each bot right now.
Pages get fetched but never cited. The crawler can read you and is choosing someone else. That is a content problem, not an access problem. Work steps three, four and six: answer-first opening sentences, real tables, and one strong page per question you actually get asked.
Only your homepage gets fetched. Assistants found your brand but not your answers. You need pages that exist to answer one question each, rather than a homepage that gestures at everything.

Fetches move before click-throughs do. If you change something this week, expect the fetch count to respond first and the referral count to lag it, because a citation has to be written before anyone can click it.

So what should you actually do

If you run a small site and have never checked whether AI crawlers can reach you, do step one today and step seven tomorrow. Those two take under an hour together and will tell you whether you have a real problem or a small one.

If your pages are already crawlable and you want the work that pays off, spend a month on steps three, four, and six. Answer-first writing, real tables, and one strong page per question you actually get asked will do more than any amount of acronym research.

If you are being asked to buy an AEO retainer and you have not yet measured anything, measure first. You cannot evaluate the work without a baseline, and the baseline is free.