SEO and Answer Engine Optimization

Ranking is not the same as being cited.

In plain English

Getting your business found, both in Google and in the AI assistants people now ask instead of searching. Those are two different jobs, and most businesses are only doing the first one.

The technical version: SEO is ranking in a list of search results. AEO, answer engine optimization, is being the source that ChatGPT, Claude, Perplexity and Google AI Overviews cite when they answer a question directly. It needs different things: content written as direct answers rather than as a build-up to one, claims specific enough to be quoted without a human editing them first, pages kept current, and a brand that is mentioned across the web rather than only on its own site. Schema and llms.txt are the cheap groundwork underneath, not the method.

What you could actually do with this

When somebody asks ChatGPT or Perplexity to recommend a person in your field, your name is in the answer.

That conversation replaces a search, and there is no page two to be on: you are either named or you are invisible.

Your pages actually appear in Google, which is not automatic and frequently is not happening.

Most sites have pages Google has never fetched even once. It costs nothing to fix and almost nobody checks.

The questions customers ask before they buy are answered on your site, in the words they used to ask them.

You stop having the same conversation over and over, and the answer works whether a person or a machine reads it.

You can see what people searched for before they contacted you.

Marketing stops being a matter of opinion, which changes what you spend on next quarter.

What it is worth

Work out what one new customer is worth to you over a year, then ask how many you would need for this to pay for itself. For most businesses the honest answer is one or two. That is the argument for doing it properly rather than cheaply, and also the reason to be suspicious of anyone promising it quickly.

What I actually deliver

The technical detail, for whoever on your side has to check it.

A technical pass: crawlability, indexing, sitemaps, canonical URLs and the reasons pages are not in the index
Content restructured to answer questions directly, in the shape answer engines extract
A plan for being mentioned off your own site: reviews, directories, local press, partner pages and communities. In a 75,000-brand study, brand mentions tracked AI Overview visibility about three times more strongly than backlinks
A refresh schedule for the pages that matter, because AI assistants cite noticeably fresher content than regular search ranks
Schema markup and llms.txt as groundwork. Cheap, easy for any machine to read, and worth having, but no rigorous study yet shows they earn citations on their own, so I do not sell them as if they do
A crawler policy that welcomes AI crawlers rather than blocking them by default, which many CDNs now do without asking
Measurement from Search Console rather than from a vendor dashboard that grades its own work

How it works

01

Find out what is actually indexed

Not what a tool estimates: what Search Console says. Sites regularly have most of their pages sitting in "discovered, currently not indexed", which no amount of content fixes.
02

Fix the technical groundwork

Sitemaps, canonicals, schema, llms.txt. This is cheap and fast. It decides whether anything else you do is even visible, but it is the floor, not the thing that gets you quoted.
03

Check you are not blocking the crawlers you need

Several CDNs now block AI crawlers by default and serve a robots.txt that forbids training. If you want to be cited, that default is the exact opposite of what you need.
04

Rewrite for extraction

Direct answers near the top, real tables, genuine question headings. An answer engine quotes a sentence, so give it sentences worth quoting.
05

Get talked about, and keep it current

A brand nobody mentions anywhere else is a brand an AI engine has little reason to name. Reviews, directories, press and partner pages build that signal, and updating the pages that matter keeps them in the running. No technical work substitutes for either.

When this is the wrong call

If your pages are not indexed, more content will not help: it joins the same queue. If the domain is brand new with no links, expect months, and treat anyone promising fast rankings as selling something else. And if the claims on the page are vague, AEO cannot help you, because an answer engine has nothing specific enough to quote. Being unquotable is a content problem, not a technical one.

Where I have done this

PracticeRank.ai

I built and run the platform that does this: automated audits, schema, llms.txt generation and content publishing straight into client CMS systems, for clients in competitive local markets.

Questions

Is AEO just SEO with a new name? +

They overlap and they are not the same. Ranking third in a list can still send you traffic. Being the source an AI cites is a different outcome with different requirements, and it is increasingly the only one that happens, because the user never sees the list.

What is llms.txt and does it do anything yet? +

A plain text file describing your site to language models, in the spirit of robots.txt. Adoption is early and no one should claim guaranteed returns from it. It costs almost nothing, it is machine-readable, and being early to a convention is cheap while being late is not.

How long before it shows up? +

Technical fixes can change indexing in weeks. Authority on a new domain takes months, and anyone quoting you a fast number for that is guessing or lying. I would rather tell you that before you pay.

How do I know whether it worked? +

Search Console impressions and indexed page counts, plus checking directly whether the answer engines cite you on the questions that matter. Both are external to me, which is the point: a vendor grading its own homework is not measurement.