AEO for Local Businesses: Getting Cited by AI

Someone just asked an AI for a plumber in your town. What did it say?

A New Way People Find You

For a long time there was one path. Someone searched, they saw ten links, they clicked one. Now a growing share of that traffic goes through an answer layer first. Google shows an AI Overview above the results. People ask ChatGPT for a recommendation. Perplexity returns a written answer with a handful of citations.

For a local business the practical question is simple. When someone asks one of these systems who to hire in your area, are you named, and is the information about you correct.

This does not replace search rankings. Traditional results still drive most clicks, and being invisible in Google will make you invisible everywhere else, because most of these systems lean on search results to build their answers. But the answer layer is a second surface, it favours slightly different things, and most local businesses have done nothing at all for it, which makes it unusually cheap to win.

What Answer Engines Are Doing

An answer engine takes a question, runs one or more searches behind the scenes, reads the top results, and writes a synthesis with citations. The important consequence is that it works by extracting claims.

So the unit of optimization changes. Traditional SEO asks whether your page can rank for a phrase. Answer engine optimization asks whether a specific sentence on your page can be lifted out, understood on its own, and attributed to you.

That difference explains most of the tactics. A page that buries the answer in the ninth paragraph of a story about your founder ranks fine and gets cited rarely, while a page that states the answer plainly, near a heading that matches the question, gets cited often.

Write The Answer, Then Explain It

The single highest value change is structural. For every question your customers ask, put a heading that is the question and a direct answer in the first two sentences underneath it. Then explain, qualify and add detail below.

Journalists call this the inverted pyramid, and it is exactly what an extraction system wants. It also happens to suit a distracted human on a phone, so you are not trading one audience for the other.

Be specific and self contained. "We open at 7am on weekdays and 9am on Saturdays" can be lifted and used. "We open early to suit our customers" cannot. Include the entity names in the sentence rather than relying on the page around it, because the sentence may be read without that context. "Emergency callouts in Bristol are available 24 hours" survives extraction. "We also do this out of hours" does not.

Numbers, prices, timeframes and named areas are the things these systems reach for most, and they are the things local sites are most reluctant to publish. Publishing a starting price is worth more citations than another page of adjectives.

Be Consistent Everywhere Else

An answer engine cross-references. If your hours differ between your website, your Google Business Profile and a directory listing, the system has three conflicting facts about you and low confidence in all of them. Low confidence means it recommends someone else.

So the groundwork is unglamorous. Name, address and phone number identical across your site, your Google Business Profile, Apple Business Connect, Bing Places and the main directories in your industry. Same formatting, same suite number, same trading name.

This is the same work as local citation cleanup, which is why the local SEO fundamentals turn out to be answer engine fundamentals too. There is no separate AEO project, just one set of facts about your business, published consistently, in a form machines can read.

Give Machines The Structured Version

Schema markup is how you state your facts in a format built for parsing rather than reading. For a brick and mortar business the ones that matter are LocalBusiness, with the specific subtype where one exists, plus Service, FAQPage and Review where relevant.

Mark up your address, geo coordinates, opening hours including exceptions, service area, accepted payment methods, and the services you offer. The point is to remove ambiguity, so that a system parsing your site does not have to infer your opening hours from a sentence.

Validate it with Google's Rich Results Test after you add it. Broken schema is common and produces no error anywhere you would naturally look.

Reputation Is An Input

These systems weight sources they can corroborate. A business with sixty reviews across Google and two industry directories, mentioned on a local news site and a supplier's partner page, is a business the model has multiple independent signals for.

A business with a website and nothing else has one signal, from a source with an obvious interest, so it gets discounted.

This is why the boring off-site work matters here. Reviews, local press, chamber and trade association listings, supplier and partner pages, sponsorships. All of it is the digital version of being known locally, and it is what makes an AI comfortable naming you.

Checking Whether It Worked

Measurement here is worse than in search, and anyone claiming precision is overstating it. There is no Search Console for ChatGPT.

What you can do is test directly. Write down the ten questions a customer would ask before hiring you, including the ones that name your town. Ask them in Google, ChatGPT, Perplexity and Gemini. Record whether you appear, whether a competitor appears, and whether what is said about you is accurate. Repeat monthly.

Then watch referral traffic in Google Analytics for AI sources, which shows up as small but usually well qualified visits. That side of it is worth a section of its own and the tracking post below covers it.

One honest caveat. For most local businesses the volume from this channel is still small next to the map pack. It is worth doing because the work overlaps almost entirely with things you should be doing anyway, and because the businesses that establish themselves as the citable answer now will be the default when the volume arrives.