How We Build Location Pages That Rank — and Get Picked Up by AI

If your customers search for what you do plus a town name, there's a new gatekeeper between you and them. Google's AI Overviews — the answer box at the top of the page — now appear on 68% of local searches, compared to 39% for the traditional map pack, according to a 2026 Whitespark study covered by Search Engine Journal. For questions with real intent behind them — "average cost of dental implants in Phoenix" — that number hits 97%.

Translation for an owner: before a customer ever sees your listing, an AI has already summarized the answer. Either your business is part of that summary, or your competitor is.

This is the process we use to build location pages that get into both — the rankings and the answer box. What we do, why, and what it looks like on a real site. (Screenshots are from client builds with names blurred — the structure is what matters.)

What is a location page?

A location page is a page on your site built for one specific market you serve — "Garage Door Repair in Cherry Hill, NJ" — with the facts, proof, and answers specific to that place. It's how a business in five towns competes in all five without pretending to be headquartered in each one.

Done right, it's the single highest-impact asset in local SEO for a small business. Done the old way, it's dead weight. Here's the difference.

Why the old location-page playbook stopped working

The old playbook was one template, duplicated per town, with the city name swapped out. Google's systems now recognize that pattern for what it is — and per the Search Engine Journal analysis, pages like that no longer generate results. Worse: publishing more thin pages without fixing their structure actively hurts you in AI retrieval, because you're teaching the machine that your domain is filler.

There's a second reason the bar moved. The research SEJ cites (an Omniscient Digital analysis of 23,000+ citations) found that your own website only accounts for about 23% of what AI Overviews cite for local queries — the rest comes from reviews, Reddit, YouTube, and directories. So the page has to do two jobs now: rank on its own, and feed the machine clean, quotable facts that match what every other source says about you.

Step 1: One real page per market — and fewer of them

We start by deleting. Most multi-location sites we audit have more location pages than they can support with real content. Ten thin pages lose to three real ones. We keep a page for every market where the business has actual jobs, reviews, and presence to show — and cut or consolidate the rest.

[SCREENSHOT 1: Anonymized locations index page, before/after — a long list of near-identical town links replaced by a shorter set of distinct market pages. Blur business name and logo.]

Step 2: Put the facts in tables, near the top

The first thing on the page after the headline is the boring stuff — and that's deliberate. Business name, address or service area, hours, and phone, formatted as a table, not prose. Per the research SEJ cites, 44.2% of AI citations come from the first 30% of a page. Machines quote what they can parse, and they parse tables cleanly.

This is also where consistency gets enforced: the name, address, and phone on this page must match your Google Business Profile, Yelp, Apple Maps, and every directory — character for character. Conflicting facts make the machine hedge, and a hedging machine doesn't cite you.

[SCREENSHOT 2: The facts table block on an anonymized location page — service area, hours, phone, license number in a clean two-column table directly under the H1.]

Step 3: Local proof — reviews and real jobs from that town

Every location page carries testimonials from customers in that market — ideally ones that name a neighborhood — and one or two real project examples with details: what the job was, what it involved, how it turned out. This is the section a template can't fake, and it's exactly the "definitive, entity-rich" material the research says AI answers prefer to quote.

[SCREENSHOT 3: Testimonial and project section — a review mentioning a specific neighborhood, next to a real project card with before/after photos. Blur names.]

Step 4: FAQs that only make sense in that market

Not "Why choose us?" Real questions with local facts in the answers: permits and township rules, climate ("Do I need impact-rated glass this close to the shore?"), seasonal timing, local pricing quirks. Three to six per page. Each answer written the way we write everything for the answer box: the answer first, in two or three sentences, then the detail.

[SCREENSHOT 4: The FAQ block on a location page — questions referencing a specific township's permit process. Blur identifying details.]

Step 5: Publish real pricing ranges

This is the one owners resist, and the one with the most data behind it: queries containing "price," "cost," or "buy" trigger AI Overviews more than 80% of the time, per the SEJ analysis. If you won't publish a range, the AI answers the pricing question with someone else's numbers.

We publish honest ranges with the variables explained — "$X–$Y depending on size and material" — either on the location page or on a dedicated cost guide the location page links to. It doesn't cost you negotiating room. It costs your competitor the citation.

[SCREENSHOT 5: A pricing-range table on an anonymized page — service tiers with ranges and a "what moves the price" column.]

Step 6: Real photos, labeled like you mean it

Stock photos tell the machine this page could be anywhere. We use photos of actual local jobs, plus recognizable local imagery, with descriptive file names and alt text that say what and where ("garage-door-installation-cherry-hill-nj.jpg", not "IMG_4032.jpg"). It's unglamorous work. It's also how a page proves it's about a real place.

[SCREENSHOT 6: The image section of a location page with the alt text/file name visible in the editor — real job photo, properly labeled. Blur the property address.]

Step 7: Structured data, then a quarterly consistency audit

Under the hood, each page gets LocalBusiness schema — the machine-readable label that says who you are, where you work, and when you're open. Then, quarterly, we audit the whole entity footprint: every directory, every profile, same name, same categories, same description. Set a reminder; drift is the default.

[SCREENSHOT 7: Schema validation result for a location page showing LocalBusiness markup passing. No business name visible in the crop.]

Step 8: Feed the 77% you don't own

Remember: roughly three-quarters of what AI Overviews cite about local businesses lives off your site. So the location-page work extends past the page — genuine review generation in each market, YouTube videos with local keywords in the descriptions, and honest participation where your customers actually talk (including local subreddits — participation, not promotion). One more move with outsized returns: publish local data nobody else has. "Average garage door replacement cost in Camden County, from our last 40 jobs" is a citation magnet, because no national site can write it.

Go deeper: Do backlinks still matter for SEO and AI Overviews? — the 27,000-domain study behind why answer quality now beats domain size, and what that means for where you spend.

[SCREENSHOT 8: Anonymized view of a review-request automation or a YouTube video description with local keywords — the off-page engine.]

What we'd update first on your site

If we opened your locations page today, the priority order is almost always the same:

1. The facts. Tables for name, service area, hours, phone — top of every location page, matching every directory.
2. Cut the thin pages. Fewer markets, covered for real.
3. Pricing ranges. The 80% stat makes this the fastest citation win on the list.
4. Local FAQs and proof. One market at a time, starting with your biggest.
5. Photos and schema. The finishing pass that makes it all machine-legible.

One honest caveat on timelines: the SEJ analysis estimates 12–18 months to build real citation authority. The facts-and-structure fixes move faster than that — but anyone promising you the AI answer box in 30 days is selling you the old playbook with a new sticker.

Go deeper: The SEO/AEO/GEO plugin for Claude Code — the tooling we run these audits with: live SERP checks, on-page crawls, and AI-visibility tracking.

Want to see how your business actually shows up — in the rankings, the map pack, and the AI answer box? Run the free Get-Found Forensic. It checks all three for your markets and shows you exactly which of these eight steps is missing. No pitch — you leave with a plan.