For twenty years, showing up in search meant ranking on a page of blue links. That page is quietly disappearing. When someone asks ChatGPT or Perplexity for the best local-first SEO tool, they do not get ten links. They get a paragraph, and your name is either in it or it is not.
Answer engine optimization, or AEO, is the practice of earning a place in that paragraph. It overlaps with SEO, but the rules are different enough that treating them as the same thing is how brands quietly vanish from the answers their buyers are reading.
You are no longer optimizing for a ranking. You are optimizing to be quotable.
The model in one sentence
An answer engine reads a question, gathers a handful of sources it trusts, and writes a synthesis. Your job is to be one of the sources it trusts, and to be legible enough that it quotes you correctly. Everything else is detail.
Those two halves matter equally. Being trusted but illegible means the model knows you exist and still describes a competitor. Being legible but untrusted means you wrote a beautiful page no engine ever pulls from.
Why the old SEO playbook only gets you halfway
Classic ranking signals still help. They are part of how a model decides what to trust. But the answer layer adds constraints that ten blue links never had:
- There is no scrolling. A model picks two or three sources, not a page of ten. Position twelve does not exist. You are cited or you are not.
- The model rewrites you. It rarely quotes verbatim. If your claim is buried in qualifiers, the synthesis loses it, and the cleaner competitor wins the sentence.
- Every engine answers differently. ChatGPT, Claude, Perplexity, and Google AI Overviews draw on different sources and weight them differently. Winning one tells you almost nothing about the others.
What legible to a model looks like
Legibility is mostly about removing friction. A model under a token budget reaches for the source that answers the question fastest and most cleanly. In practice that means:
- A direct, declarative answer near the top of the page instead of buried after three paragraphs of preamble.
- Claims that stand on their own when lifted out of context, with the specifics attached such as numbers, dates, and named methods.
- Structure a parser can follow: clear headings, real FAQs, and schema that says plainly what the page is about.
None of this is a trick. It is the same thing that makes a page useful to a human in a hurry, which is rather the point.
Where to actually start
You do not need a new department. You need to know, prompt by prompt, where you currently stand and then fix the two or three answers that matter most to your business. Pick the questions a buyer would really ask, see who the engines cite today, and work backwards from the gap.
The honest version of this work is unglamorous and measurable. You appear in 42% of tracked answers. A competitor appears in 61%. Here is the one page standing between you. That is a problem you can solve. Do more AEO is not.
The one rule we will not break
It is tempting, in a field this new, to invent confidence, to claim a citation you did not verify or a result you cannot see. We do not, and you should not either. If an engine is not showing your data, the truthful answer is we do not know yet, not a plausible guess. Optimizing for answers only works if the answers are real.
This is the loop the app runs. It watches which engines cite you, flags the slips, and drafts the brief to make a page more quotable locally on your machine before anything goes live.
Next up, we will get concrete about how to write a single brief that turns one of those lost answers back in your favor. Until then, go look at where you actually stand.
Spent a decade doing SEO and AEO in the trenches, then built the local-first tool he always wanted. Named the company after his dog.