How AI Overviews Are Changing Local Search

The map pack did not disappear. That is the first thing to settle, because a lot of advice written in the past year assumed it would.

What happened instead is subtler and harder to plan around. On some local queries Google now writes an answer above everything else, sourced from places you may not control, and the searcher forms an opinion before they ever reach the three business listings underneath.

If you run a local business, the useful question is not whether AI Overviews are good or bad. It is which of your queries trigger one, where the answer is coming from, and which of those sources you can actually influence. Those have concrete answers. Our general explainer on what Google AI Overview is covers the mechanics across all query types; this covers the local layer specifically.

Local Queries Trigger Overviews Less Often, and Not Where You Expect

Google has been noticeably more conservative with AI Overviews on local intent than on informational intent, and the pattern is fairly consistent.

A bare transactional query with obvious local intent, the "plumber near me" shape, usually still returns the map pack with no overview above it. Google appears to treat these as already answered by the local results, and an overview adds nothing.

The queries that do trigger overviews are the ones with a research shape attached to a place:

• "best roofing company in Katy"
• "how much does a plumber cost in Toronto"
• "do I need a permit to replace a deck in Brampton"
• "is a tankless water heater worth it in Florida"
• "what to look for in an HVAC contractor"

These are the queries a customer runs before the "near me" search. That timing is the whole point. The overview is now shaping the shortlist that later becomes a map pack click.

The Overview Sits Above the Map Pack, Not Instead of It

On a local query that does trigger one, the usual layout puts the AI Overview first, the map pack under it, then organic results.

Nothing was removed. Everything moved down. The practical effect is that your map pack position is now competing for attention with a block of text that has already partly answered the question, and on a phone that block can occupy most of the first screen.

This is why the "we still rank number one locally" reassurance is worth checking rather than trusting. Ranking first in a pack that now sits below the fold on mobile is a different commercial position than it was two years ago, even though the ranking report reads identically.

Worth measuring from the places your customers actually are. Our Local Search Tool attaches a location parameter to a search link, so you can see whether your research-stage queries trigger an overview in the areas you serve, and what it says. Results vary by location more than most people expect.

Where the Answer Comes From

For local queries, Google draws on a mix of sources, and the balance is different from informational queries.

Google Business Profile data is used directly. Hours, categories, attributes, services and the business description all feed answers about what you do and whether you are open. Incomplete profiles produce vaguer mentions.

Reviews are used as source material rather than only as a ranking signal. This is the change most local businesses have not adjusted to, and it is covered further down.

Third-party lists and directories are cited heavily. For "best X in [city]" queries, the sources under an overview are frequently a local publication's roundup, a directory category page, or a national aggregator, rather than the websites of the businesses being discussed.

Your own site is used mostly for specifics: pricing, process, service area, licensing, what a job involves. Pages that answer a defined question get pulled in. Pages that describe how committed your team is do not.

Want to know which of your local queries now return an AI answer before your listing? Get a Free Visibility Audit →

"Best in City" Became a List-Publisher Game

This deserves its own point because it changes where the work goes.

When someone searches "best electrician in Sugar Land", the overview usually synthesises from pages that rank businesses, not from the businesses themselves. That is a structural preference: Google wants a source that compares options, and your own site cannot credibly be that source.

So the work moves to getting into those lists. Local publication roundups, chamber directories, trade association member listings, the established directories in your vertical, and any "top ten" page that already ranks for the term. These have always been worth having and are now worth considerably more, because they feed the answer rather than sitting on page one where nobody looks.

Getting listed is unglamorous work: contacting the publication, meeting whatever criteria the directory has, keeping the listing accurate. It is also one of the few parts of this you can directly cause to happen.

The Questions Worth Owning on Your Own Site

If the "best in city" comparison is largely out of reach, the specification questions are not. Your own pages are cited well for anything concrete enough to have a definite answer, and those questions are the ones customers ask immediately before hiring.

Four categories work consistently for local businesses.

1. Cost. What a job actually runs in your market, what moves the number up or down, and what a quote includes. Most competitors refuse to publish this, which is exactly why the pages that do get pulled into answers.
2. Process and timeline. What happens on day one, how long the work takes, what the customer needs to arrange. Specific enough to be quotable.
3. Permits and local requirements. What a job needs in your city, who pulls the permit, roughly what it costs and how long approval takes. This is genuinely local, it changes between municipalities, and a national competitor cannot write it.
4. Comparisons within your own category. Repair against replacement, one material against another, the cheaper option against the one you would install in your own house. Say which you would choose and why.

The format matters as much as the topic. Put the direct answer in the first two sentences under a heading that matches how people ask the question, then expand underneath. Answers get extracted from passages, so a page that buries the number in paragraph nine gets skipped in favour of one that leads with it.

Say the awkward part too. A page that admits when the cheaper option is fine, or when a customer does not need the service at all, tends to get cited more than one that recommends the expensive path every time. It also converts better with the people who do need you.

Reviews Are Now Source Text

Google has been summarising review sentiment for a while, and AI Overviews extend it. Answers about local businesses increasingly reflect what reviews say, in the language reviewers used.

Two consequences follow, and both are actionable.

The specific words in your reviews matter more than they used to. A business with forty reviews saying "showed up on time" is described differently from one with forty reviews saying "great service". You cannot script reviews, and attempting to is against Google's policies and obvious to readers. What you can do is ask at the moment the specific thing went well, because people describe what is fresh.

Recurring complaints now compound. One review mentioning a missed appointment is noise. Six of them become a characteristic that an AI answer can state as fact to someone who never scrolled to read them. The mechanics of how reviews feed rankings are in how reviews affect Google rankings; the new part is that they also feed the sentence a customer reads first.

Entity Consistency Decides Whether You Are Mentioned at All

For an AI answer to name your business, Google has to be confident about which business it is. That confidence comes from consistency across your profile, your site, your structured data and the places you are listed.

Name, address and phone consistency has been a local SEO fundamental for years, and the payoff has changed rather than the task. It used to affect ranking. Now it also affects whether a generative answer is willing to name you specifically or falls back on a generic description of the category.

Check what Google already holds. Our Knowledge Graph Search queries Google's entity data for a business name and shows what comes back, which is often clarifying. If Google holds no entity for you, or holds one with the wrong details, that is the first thing to fix. Structured data helps here, and Google Search Central's local business markup guidance covers the fields that matter.

What Happens to Your Clicks

Some queries that used to produce a visit now end in the overview. That is real, and pretending otherwise helps nobody.

The offsetting factor for local businesses is that local intent converts through actions that are not website clicks. Calls, direction requests and messages from the profile continue regardless of what the overview says, and a customer who read a favourable answer and then tapped call is a better lead than one who browsed three sites first.

Which means the reporting needs adjusting. If you judge local performance by sessions alone, you will read a shift in behaviour as a decline. Track calls, direction requests and profile interactions alongside traffic, and compare the trend across all of them before concluding anything.

Measuring It Honestly

There is no clean report for this, and anyone offering you one is selling something.

Search Console does not separate AI Overview appearances from other impressions, so you cannot isolate them there. What works is unglamorous: build a list of the twenty or thirty research-stage queries that matter in your market, check them manually from your actual service areas on a set schedule, and record whether an overview appears, what it says, and which sources it cites.

That takes an hour a month and tells you more than any tool currently on the market. It also catches the case that matters most, which is an overview citing a competitor or repeating something inaccurate about your business.

The Part Nobody Can Promise

There is no method that guarantees inclusion in an AI Overview, and the outputs change without notice. A query that returns an overview today may not next month, and the sources cited shift.

What holds up is the underlying position: a complete and accurate profile, presence in the local lists that get cited, reviews that describe specific things you did well, consistent entity data, and pages that answer the questions customers actually ask before they buy. Every one of those was worth doing before AI Overviews existed and pays off in the map pack regardless of what the overview does.

That is the honest version. Treat anyone promising AI Overview placement the way you would treat someone promising a specific Google ranking. If you want a structured way to work through the wider picture, our AI Engine Optimization service covers how these answers get built across Google, ChatGPT and the rest.

Not sure what AI answers are saying about your business in the places you serve?

A Free Visibility Audit covers which of your key local queries return an AI Overview, what those answers say, which sources they cite, and where your profile and entity data are letting you down.

Get a Free Growth Audit →

---

Related reading: What Is Google AI Overview? | How to Dominate Google Maps | What Is AI Engine Optimization?