The Home Service Business Guide to AI Search Visibility

A homeowner with a leaking water heater used to type a search and pick from ten blue links. Now many of them ask ChatGPT, or read a Google AI Overview that names a few businesses and stops there. If your business isn't one of the ones named, you don't see it as a lower ranking. You just see fewer calls.

This guide is the one place that pulls together what to fix first if you want AI to name you: where these tools get their information, the specific things you can change, and what's still genuinely unsettled. Two companion posts go deeper on the two most technical pieces, what actually matters for home service AI visibility and the specific fixes for getting cited. Read this one first.

In short: there's no separate "AI SEO" system to break into. Google's own documentation says you don't need special markup or files to appear in its AI features (AI features and your website, Google Search Central). The same four things that help you in the map pack, your profile, your schema, your consistency across sources, and content that answers a real question, are the sensible place to start for AI answers too. The rest of this guide walks through each one, plus the two places where AI visibility may pull apart from traditional local search.

Where AI Assistants Get Their Information

Google says that to be shown as a supporting link in AI Overviews or AI Mode, a page has to be indexed and eligible to appear in Google Search with a snippet (AI features and your website, Google Search Central). ChatGPT and Perplexity can search the web when they answer, so the same public pages matter for them: your site, your Business Profile, directories, review platforms, local news, anywhere your business gets talked about.

Which of those sources a given tool leans on most isn't something the tools publish, and we haven't tested it properly yet. Treat any claim that one engine prefers one source as a guess, including any you read elsewhere.

That has one practical consequence worth sitting with: you're not optimizing for a new system. You're optimizing the same public information these tools can already read, so that whatever is synthesizing an answer can extract a clean fact from it instead of skipping past it.

The Four Things Worth Fixing First

Strip away the vendor jargon and it comes down to four things. They're listed in the order we'd check them, not in order of proven impact, because we don't have data that ranks them for AI answers:

• Google Business Profile accuracy and activity. Correct categories, current hours, a steady stream of real photos, and posts that don't go stale. A profile still listing "General Contractor" as the primary category when you've specialized in roofing for five years gives every system that reads it a vaguer picture of your business than the one you'd want. It's also usually the quickest of the four to fix, because it's a settings change you control directly.
• Structured data on your site. LocalBusiness, Service, and FAQ schema turn your prose into a machine-readable version of the same facts. Google says none of it is required for its AI features, so think of it as clear labeling, not a lever.
• Consistency across every source. Your site, your profile, directories, and reviews need to describe your services and area identically.
• Content that answers a real question. "How much does a typical water heater replacement cost in [city]" is answerable. "Your trusted local plumbing partner" is not. AI tools are built to extract answers, not slogans.

Consistency is the one people underestimate, and it's the cheapest to check. If your website says you serve three towns and your Google profile says five, a person has to guess which is right, and so does a tool. Making the sources match removes the question. We haven't measured how much this moves AI answers, so we're treating it as good hygiene with a plausible upside, not a proven lever.

What to Skip

A few tactics that get sold as AI SEO have no evidence behind them that we've seen, and it's worth naming them so you don't spend budget on the wrong fix.

Bulk directory listings and bought backlinks are the first. Nothing we've seen shows them helping with AI answers, and a pile of paid listings that all repeat the same text isn't independent confirmation of anything. Keyword-stuffing your service pages is the second. Google's Danny Sullivan has said that optimizing for AI answers is part of SEO and comes down to writing for people, not ranking systems (Google's Danny Sullivan: SEO for AI is still SEO, Search Engine Land, December 2025), and repeating a phrase doesn't make a fact any easier to state. The third is any "AI SEO certification" that promises a place in a preferred index. We don't know of any such index, so be skeptical of anyone selling access to one.

Structured Data Is Infrastructure, Not a Shortcut

It's worth being straight about what schema markup does, because a lot of what gets sold as "AI SEO" oversells it. Google's documentation says you don't need new machine-readable files, AI text files, or markup to appear in AI Overviews or AI Mode, and that there's no special schema.org structured data to add. It also recommends that your structured data match the visible text on the page (AI features and your website, Google Search Central). Sullivan has made a similar point in public: GEO is part of SEO, not a separate discipline. So schema is plumbing. It doesn't replace having a real, specific answer on the page; it just makes that answer easier for a machine to read correctly.

For a home service business, three schema types are a sensible starting point: LocalBusiness (name, address, phone, service area, hours), Service (one entry per trade you offer), and FAQ on any page where you're genuinely answering real customer questions. We're covering what to add and where it goes in a separate guide. For now, the sequence that matters is: write the plain-language fact first, then wrap it in schema, and keep the schema matching what's visible on the page. Schema around a page with no real answer on it doesn't fix the page.

AI Overviews and the Map Pack Are the Same Project

It's tempting to treat "getting into AI Overviews" as a new, separate initiative from "ranking in the map pack." Don't. The two draw on overlapping public information: your Business Profile, your reviews, and the content on your site. Where they likely differ is in what each one needs from your content. The map pack leans on profile signals and proximity, while an AI answer needs a sentence it can state. We haven't measured how often the two diverge, so read the example below as an illustration, not a finding.

The practical takeaway: if you're already doing the work to rank in the map pack, GBP cleanup, review generation, service-area pages, you've done much of the same work. This isn't a second checklist to run. It's the same checklist with one more filter on top: does this fact stand on its own as a sentence a model could lift directly.

Here's a hypothetical that shows how the two could pull apart. Picture a roofing company with a dozen strong reviews and a clean profile but a bare-bones website. It could still hold a solid map-pack position on local relevance signals alone. Ask an AI assistant to name a roofer in that same town, though, and it may skip that business entirely, because there's no page anywhere stating plainly what the company specializes in or what a typical job costs. In that scenario the map pack rewards proximity and reviews, and the AI answer needs a sentence to quote and doesn't find one.

What We're Testing Next: Real Queries, Real Answers

Much of what's written on this topic, including parts of this guide, describes the mechanism in the abstract. Few write-ups we've seen show an actual exchange: a real query typed into ChatGPT or Perplexity, the real answer that came back, and which businesses got named. We run automated audits against real contractor sites and profiles for a living, which puts us in a reasonable position to actually run that test properly instead of describing it secondhand.

The test we're running asks the plain questions a homeowner actually types: "who's a good emergency plumber near [city]," "best rated roofer in [city] for a full replacement," "HVAC company near me that does same-day repairs." We're logging which businesses get named across ChatGPT, Perplexity, and AI Overviews for the same query set, and what those businesses have in common on their profile, their reviews, and their site.

That post isn't finished yet, and we're not going to publish a number we didn't verify just to fill this section. When it's done, it'll show the real query pairs, the actual answers, and what separated the businesses that got named from the ones that didn't. Check back, or better, run the free audit now and we'll tell you specifically where your business stands on this rather than in the aggregate.

The Review Language That Gets Quoted Back

A star rating gives a tool nothing to quote. A review that says "fixed our furnace on a Sunday within two hours of calling" hands it a sentence it could lift directly into an answer. This part of AI visibility depends less on your website and more on how and when you ask for reviews. We haven't yet measured how often review text shows up in AI answers, and it's one of the things the query test above is meant to check.

Timing and specificity both matter. Asking right after the job, while the details are still fresh, tends to produce language a customer actually remembers, not a generic "great service, would recommend." Specific language like that gives a tool something concrete to quote, where a generic line gives it nothing. It's the same asset your review platform profile and your site should both be surfacing consistently. We're covering how to ask for reviews in a separate guide on reviews and reputation.

Do You Need an Agency, or Can You Fix Most of This Yourself

Most of this you can do yourself this week:

• Log into your Google Business Profile and check every category, hour, and service listed against what's actually true today.
• Add ten real photos of finished jobs and your crew if your profile is running on old or borrowed images.
• Pick your three most-asked customer questions and write a direct, specific answer to each on your site, with your service area and a real price range stated plainly.
• Check that your service area and offerings read identically on your site, your profile, and your two or three biggest directory listings.

None of that requires a vendor, and it's a reasonable afternoon's work.

Where it tends to get harder without help is schema implementation done correctly across a whole site, and staying consistent across every directory and platform as your business grows or changes. That's less a skill gap than a time-and-attention gap, and it's a reasonable place to pay someone rather than do it yourself.

If you're not sure which bucket your business falls into, that's exactly what the free audit is for. It checks your site, your Google Business Profile, your review standing, and whether AI assistants currently name your business, and tells you plainly which of these four levers is actually holding you back.

The Four Levers, Summarized

AI visibility is a citation game, not a ranking game. Four things are worth fixing first: your profile, your schema, your consistency across sources, and whether your content answers a real question instead of selling one. Most of it builds directly on work you're probably already doing, or should be, for the map pack. The parts that are genuinely new, mostly the review language and the structured data, are covered in depth in the two posts linked above and the ones still coming. Start with the free audit if you want to know which of the four is costing you the most right now.