Google AI Overviews Aren't Showing Your Business. Here's the Fix.

The Disconnect: Rank #1 in Map Pack, Invisible in AI Overview

We audit roughly 200 local businesses a year. One pattern shows up over and over. A business ranks in the top 3 of the Map Pack for their core query. Their Google Business Profile is dialed in. They have 200+ reviews averaging 4.8 stars. They have done everything Google's local guidelines tell them to do.

Then you run the same query and the AI Overview cites three other businesses. Often businesses with worse reviews, smaller GBP profiles, and lower Map Pack positions. The owner asks the obvious question: how am I ranked #1 in the Map Pack but invisible in the AI Overview that sits above it?

The answer is that the Map Pack and the AI Overview are two different products from Google running on two different signal mixes. They share inputs, but the weighting is fundamentally different. Map Pack is proximity-heavy and review-count-heavy. AI Overviews is content-heavy, schema-heavy, and freshness-heavy.

If you have not seen the cross-engine pattern before, we wrote about it in the Google vs ChatGPT local search breakdown. Same logic applies inside Google itself. The Map Pack and AI Overview disagree because they are answering slightly different questions.

What AI Overviews Are Pulling From (Hint: Not Just GBP)

Based on Google's own documentation, observed citation behavior across 15,000+ queries we have tracked, and patent filings around the Search Generative Experience, AI Overviews appears to weight signals roughly in this order for local queries:

1. Structured data on the source page. LocalBusiness schema with the correct industry subtype, FAQPage schema, Service schema, and Review schema. Pages with clean structured data get cited at 3-4x the rate of pages without.
2. Semantic relevance of page content to the query intent. Does the page actually answer the question being asked, or does it just list services? Decision-stage content wins here.
3. Freshness signals. Pages updated in the last 90-180 days outperform stale pages. Datestamps, recent content additions, and last-modified headers all matter.
4. Site authority and topical depth. A site with 12 related blog posts on a topic gets cited more than a site with one. Internal linking matters.
5. GBP signals as verification. Categories, hours, NAP, photos. Used to confirm legitimacy, not as a primary ranking input.
6. Review count and rating. A meaningful signal, but ranked lower than the Map Pack ranks it. We have seen 50-review businesses cited over 300-review businesses on the same query.

The implication is uncomfortable for businesses that have spent years optimizing GBP and review velocity. Those signals still matter for the Map Pack, but they are not enough for AI Overviews. The full GBP optimization guide for AI covers the GBP-side fixes, but the bigger lift is on-page.

The 3 Changes That Flip Most Local Businesses

In our last 40 AI Overviews recovery engagements, three changes accounted for roughly 80% of the lift. Not a long checklist. Three specific moves.

Change 1: Industry-Specific Schema, Not Generic LocalBusiness

Most local businesses either have no schema or have generic LocalBusiness schema. AI Overviews wants industry subtype. If you are a dentist, the @type is Dentist, not LocalBusiness. If you are an HVAC company, it is HVACBusiness. If you are a law firm, it is LegalService or Attorney. The list of valid subtypes lives in Schema.org and Google's structured data documentation.

Layer on FAQPage schema with real questions (the same questions appearing in People Also Ask for your core query) and Service schema for each individual service. Three schemas, properly nested, on every service page. We covered the full implementation in the schema markup guide. Most developers can ship this in 2-4 hours per template.

Change 2: Decision-Stage Content, Not Service-List Pages

Service pages that just list what you do are informational. AI Overviews cites decision-stage pages, the ones that help a buyer decide. Examples: pricing transparency posts, "how to choose a [provider] in [city]" guides, comparison pages between your service and an alternative, real-cost breakdowns, and what-to-expect timelines.

The pattern: the page must directly answer a buying-stage question that a real customer is typing into Google. Not "our services" or "why choose us." Real questions like "how much does a root canal cost in Phoenix" or "what should I look for in an emergency restoration company." Those queries trigger AI Overviews more reliably, and pages built specifically to answer them get cited.

Change 3: 90-Day Refresh Cadence

AI Overviews favors fresh content more aggressively than the Map Pack does. Every key page on your site (homepage, service pages, top decision-stage posts) should have a visible last-updated date and a fresh paragraph added every 90 days. The fresh paragraph should reference current data, recent changes in your industry, or new questions customers are asking.

The mechanism: AI Overviews appears to discount pages that have not been touched in 12+ months, even when those pages are technically still accurate. A 200-word update every quarter signals to the engine that the page is being maintained, which improves citation rate.

Why Schema Markup Matters Way More Here

Schema is the load-bearing wall. The reason: AI Overviews has to summarize quickly and accurately, and structured data is the fastest, cleanest way for the engine to confirm what a business is, what it offers, and what other entities (services, locations, FAQs) connect to it. A page without schema forces the engine to infer everything from unstructured text. A page with clean, layered schema hands the engine the answer.

In our last citation rate audit across 200 local sites, pages with industry-specific LocalBusiness + FAQPage + Service schema got cited in AI Overviews at 18.4% of tracked queries. Pages with only generic LocalBusiness got cited at 4.2%. Pages with no schema got cited at 0.9%. The schema delta alone is a 20x lift.

The implementation checklist:

• LocalBusiness with the correct industry subtype (Dentist, HVACBusiness, LegalService, MedicalBusiness, etc).
• FAQPage on every service page with 6-10 real PAA questions pulled from your core queries.
• Service schema for each service offering with provider, areaServed, and serviceType filled out.
• Review schema (aggregateRating + sample reviews) if you have public reviews and the page they live on.
• BreadcrumbList on every interior page so the engine understands site hierarchy.

Validate everything with Google's Rich Results Test before going live. Schema errors silently break citation eligibility, and the tool catches them fast.

How to See What AI Overview Is Showing for Your Queries

Three methods, ordered by reliability.

Method 1: Manual incognito queries. Build a list of 15-20 high-intent queries for your service in your city. Run each one in an incognito browser every two weeks. Screenshot whether an AI Overview appears and whether you are cited. Log the results in a simple sheet. This is the most reliable data because it is what your actual customers see.

Method 2: Google Search Console. GSC now reports AI Overview impressions under the Performance report. Filter by Search Appearance and look for the AI Overviews row. Coverage is incomplete and there is a 2-3 day lag, but it is first-party data and useful for trend lines.

Method 3: Third-party tools. SE Ranking, AlsoAsked, and Profound now track AI Overview presence at the query level. They are imperfect and they sample, but they give you trend lines without manual screenshotting. We use them for client reporting, never as the only data source.

A 30-Day Action Plan

If you have read this far and want to actually fix the gap, here is the 30-day plan we run with clients. No fluff, just the order of operations.

Week 1: Baseline and schema audit.Run 15-20 incognito queries, screenshot every AI Overview, log who is cited. Audit your existing schema with Google's Rich Results Test. Identify which industry subtype you should be using. Pull your competitors' schema for comparison.

Week 2: Schema implementation. Ship industry-specific LocalBusiness, FAQPage, and Service schema to every service page. Validate each one. Re-submit the pages to Google Search Console for re-indexing. Most of this is 6-10 hours of developer work.

Week 3: Decision-stage content. Pick two high-intent, AI-Overview-activating queries from your tracking list. Write one page for each that directly answers the query. Pricing transparency, comparison, or how-to-choose format. 1,200-1,800 words each. Publish with full schema implementation.

Week 4: Freshness pass and re-measurement. Add a last-updated date and a fresh 150-word paragraph to your top 5 existing pages. Re-run the 15-20 baseline queries. Compare to week 1 screenshots. Most clients see early movement on 2-4 queries by day 30, with full citation rate lift continuing into days 45-60.

Frequently Asked Questions