What is Answer Engine Optimization?

Search gave you ten chances to be chosen. An answer gives you one paragraph naming two or three businesses. There is no page two to grind up from.

From ten results to one answer

Search used to give you ten chances to be chosen. An answer engine gives you one paragraph naming two or three businesses. Either you are in it or you are invisible — there is no page two to grind your way up from.

Answer engine optimisation is the practice of making sure the systems writing those paragraphs have accurate, well-sourced, easily parsed information about you. It overlaps with SEO and is not the same job. Ranking is about beating other pages. This is about being described correctly by something that will never show your page at all.

Why this happened now

Nothing about your business changed. The place people ask changed.

For twenty-five years the interface to local information was a list of links, and an entire industry grew up around getting onto that list. Assistants replaced the list with a sentence, and a sentence has room for two or three names. The underlying question — who should I call — is exactly the same question people were typing into search boxes, arriving at a surface that answers it differently.

That is worth stating because it sets expectations. This is not a new marketing channel to be conquered with new tactics. It is the same demand, met by an interface with far less room in it, which is why the consequences of being described inconsistently are so much harsher than they used to be.

Where the answers actually come from

The common assumption is that an assistant knows things about your business. For a local business it almost never does — there was never enough written about you for a model to have learned you as a fact.

What happens instead is that it searches, reads a handful of pages, and writes from those. Directories, review platforms, an aggregator for your trade, occasionally a community thread, and your own site if it has anything relevant to say. Then it reconciles those accounts into one description.

That single mechanism explains almost everything else. It is why a correction on your website does not fix an answer, why the same question can produce different answers, and why the practical work is unglamorous. What gets read, in detail, is worth understanding before you spend money on any of this.

The two halves of the job

Almost every piece of advice in this subject is one of two things, and separating them makes the whole discipline tractable.

Agreement. Making the facts about you identical everywhere they appear. Name, address, phone, hours, categories, towns. Mostly performed on other people’s websites, mostly dull, and it decides whether you are described confidently or hedged into vagueness.

Substance. Having something worth quoting that only you can supply. What the work costs, which towns you cover, what a specific job involves. Performed entirely on your own site, and it decides whether you are the answer to a particular question or merely a business that exists.

Do the first without the second and you are accurate and forgettable. Do the second without the first and your careful new pages get outvoted by records you never claimed — which is the more common and more demoralising failure.

What the work actually consists of

Three things do most of the lifting, and none of them is exotic.

• Consistency across the sources engines cite. Your name, address, hours and service list should say the same thing everywhere they appear. Where they disagree, the engine picks — usually not the one you would have picked.
• Plain language on your own site. State what you do, where you do it, and who for, in sentences. Not in a hero image, not in a carousel, not implied by photographs of vans.
• Some evidence you exist. Reviews, directory listings, a mention in a local thread. An engine that finds nothing about you writes about somebody else.

What it is not

It is not a submission process. There is no form, no directory of businesses that assistants consult, and no verification badge. Anybody offering to add you is selling something else under that name.

It is not placement you can buy. Nobody sells a spot in an answer today. If that ever changes it will be labelled advertising, the way search ads are.

It is not a trick. There is no markup you can add, no phrase you can bury, no instruction you can hide in your source that makes an engine name you. Those all either do nothing or actively count against the page.

It is not entirely new. A great deal of competent local search work carries over untouched, and the honest comparison matters more than the acronym.

The order that actually works

Measure, agree, write, then plumbing. In that order, because each stage makes the next one worth more.

Measure first. Ask the questions your customers ask, in your town, without naming your business — naming it manufactures the mention. Record who was named and what was cited. Fifteen minutes, and it tells you which of the following is actually your problem.

Then fix the records. Claim the listings that appeared in those citations and make them identical rather than similar. Start with the sources that keep coming back; if a specific fact is being repeated wrongly, hours are the usual culprit.

Then write the missing facts. Prices, or a range. The towns you cover, as names. A page per service rather than one page listing everything. This is the half that gets your own site cited instead of summarised out of existence.

Then the plumbing. Check nothing is blocking the crawlers, publish the structured data, and add an llms.txt. An afternoon, and it only matters once there is something worth reading.

How you would know whether it worked

Not by checking once and screenshotting the result. The same question genuinely returns different answers between runs, so a single check is one draw from a distribution rather than a reading.

What can be measured is a rate: across a fixed set of realistic questions, asked repeatedly from a fixed location, how often were you named, and how often was a rival named instead. That is what an honest score is built from, and it is a different instrument to the ranking reports you may be used to.

It is also worth knowing what such a measurement cannot tell you before you buy one from anybody, us included.

The vocabulary, briefly

You will meet a few terms, and they are simpler than they sound.

Answer engine — anything that responds to a question with prose rather than a list of links. That includes the standalone assistants and the generated answers now sitting above ordinary search results.

Citation — a source the engine drew on and is willing to show you. These are the most useful artefact in the whole subject, because they are both the evidence and the thing you can go and correct.

Grounding — the step where the engine searches and reads before writing, rather than answering from memory. Local answers are almost entirely grounded, which is why current records matter more than anything a model absorbed in training.

Retrieval — the same idea from the other side: which pages came back and got read. Being in that small set is the whole game.

None of these needs to be in your vocabulary to do the work. They are here so that when a supplier uses one, you know whether they are describing something real.

The uncomfortable part

None of this is fast, and some of it is not under your control. You cannot make a directory correct its record today, and you cannot make a community thread mention you at all.

Nor can anybody promise you a place in an answer. The engines are not obliged to name any business, they change how they answer without notice, and a guarantee would require control that no vendor has.

What you can do is find out what is wrong before it costs you a customer, and fix it in the order that matters. That ordering is the whole job — and it starts with knowing which sources the engines are actually reading. The cheapest first move is usually an llms.txt file, because it is the one input you control completely. If somebody is about to sell you a visibility number for this, read what one is actually made of first.

Who it does not help much

Worth saying, because the honest version of this is not that everybody needs it equally.

If almost all your work comes through repeat customers and personal recommendation, and you are not trying to grow, this will change very little for you. If you are in a category where people do not research before choosing — genuine emergencies where the first answered phone wins — the research half matters much less, though being named still does.

And if your problem is that customers who find you do not book, this is the wrong diagnosis entirely. Visibility work makes more people consider you. It does nothing about what happens after that.

Is it worth doing at all?

For most local businesses, yes, and for a duller reason than the marketing suggests.

The work is finite. The list of sources that describe you is short and knowable, and once they agree, they largely stay agreed. The pages you write to state your prices and your coverage do not need rewriting quarterly. This is not a channel you feed forever; it is a set of repairs and a handful of pages.

A caution to go with that. Because the work is finite, it is also easy to oversell — a retainer implies ongoing labour, and once the records agree there is not a great deal of ongoing labour to do. Reviewing the sources a couple of times a year and writing a page when your services change is genuinely most of the maintenance.

And most of it pays twice. Correct records and a site that states what things cost are good for the customers who find you by any route, including the ones who never ask an assistant anything. That is the honest case for it: not that a new technology demands a new budget, but that the things it rewards were worth doing anyway, and now there is a reason to actually do them.