Arif Rahman

How Google’s AI Overviews Are Changing What Shows Up From Your GBP

Key takeaway: AI Overviews are no longer just a summary box sitting politely above your search results. For local queries, they are increasingly displacing the traditional 3-pack itself, showing one or two businesses instead of three, dropping the call button, and pulling from your Google Business Profile using selection logic that overlaps with but isn’t identical to local pack ranking. Your GBP hasn’t stopped mattering. It has become the raw material for a system that decides whether you get mentioned at all.

Most advice treats AI Overviews like a new layer stacked on top of search, something that sits above the local pack and organic results without touching how they work underneath. That premise doesn’t hold for local search anymore. In a growing share of “near me” and service-category queries, the AI Overview isn’t sitting above the local pack. It’s replacing it.

The 3-pack isn’t just getting company. In some queries, it’s getting replaced.

The traditional Google local pack is a familiar object: three businesses, a small map, star ratings, a phone number you can tap. It’s been ranked the same way for years, primarily on proximity, relevance, and prominence, with your Business Profile as the main data source. Google has now started rendering an AI-generated local result in place of that widget for a meaningful and growing slice of queries. Here’s what actually changes when that happens:

Traditional 3-packAI-generated local result
Businesses shown3Usually 1–2
Call buttonYesOften missing
Selection logicRanks a shortlist by proximity, relevance, prominenceConstructs an answer, then decides who to name in it
What happens instead of a callTap-to-callClick into the AI’s pick, or an ad

Local SEO analyst Joy Hawkins has been tracking this shift closely through 2026, and the pattern she’s flagged is consistent with the table above: some businesses she’s tracked have seen visibility drops of 50% or more on the queries affected. Search Engine Roundtable’s coverage of her findings is one of the more grounded, source-traceable accounts of this rollout, in a topic otherwise flooded with vague marketing-blog statistics.

That last part is worth naming directly, because the numbers floating around don’t agree with each other:

Claimed statRange across sources
AI Overviews appear on X% of all searches15% to 68%, depending on the source
AI-powered local packs appear on X% of local queriesroughly 8%, per one tracked source
GBP’s weight in local pack ranking32% to 36%, depending on the source

These aren’t necessarily wrong so much as unverifiable. Different tools measure different query sets at different times, and a lot of this content reads like it was written to fill a keyword gap rather than researched. The honest position: AI Overviews are expanding into local search, the rollout has been active since late 2025 and through 2026, and the effect on visibility is real and documented. The exact percentage isn’t something to build a strategy on.

Your GBP feeds the answer. It doesn’t get consulted the same way twice.

Here’s the mechanism worth understanding, because it changes how you should think about optimization.

The local pack and AI Overviews both draw from the same underlying source: your Google Business Profile, the Knowledge Graph entity built around your business, and third-party signals like reviews and citations. Google hasn’t built a parallel database for AI answers. It’s using the same profile data you already manage.

What’s different is the selection and packaging logic sitting on top of that data. The local pack ranks businesses and shows a fixed number of them, generally three, based on proximity, relevance, and prominence. An AI Overview doesn’t rank a shortlist. It constructs a written answer to the specific question implied by the search, then decides which one or two businesses to name as part of that answer. A search for “plumber near me” triggers ranking logic. A search for “best plumber for emergency water heater repair near me” triggers something closer to synthesis, where Google’s model is trying to answer a more specific implied question, and your profile either supplies the specific detail that answers it or it doesn’t.

This is why a business can rank respectably in the 3-pack and still get skipped in the AI Overview for a more specific version of the same query. Ranking and citation are related but separate outcomes.

What actually determines whether your business gets named

Strip away the inflated stats, and the same four signals keep showing up across credible reporting on this and Google’s own documentation of how AI Overviews source information:

SignalWhat it looks like done wellWhy it matters here
Profile completeness and specificityCorrect primary category, filled-out service listings, a description with real specifics (what you do, for whom, where)Gives the AI concrete text to extract instead of generic filler
Review recency and sentimentFresh reviews in the last few months, not just a high star ratingRecency signals an active business; sentiment and specific mentions (wait times, staff, services) become quotable material
NAP consistency and structured dataName, address, phone match across GBP, website, and directories; Local Business schema matches GBP dataReduces ambiguity; mismatched data can push Google toward a competitor with cleaner data
Active, current engagementRecent posts and photos, not a profile last touched months agoA freshness signal that tells Google the business is still operating and worth surfacing

None of these are new local SEO advice dressed up for an AI news cycle. What’s new is that they now feed two different selection mechanisms at once, so neglecting them costs you visibility in both places rather than one.

The real cost isn’t just a lower rank. It’s fewer seats at the table.

When an AI Overview replaces the 3-pack on a query, the businesses affected aren’t just reshuffled, they’re often cut from three slots down to one or two. There’s frequently no call button in the AI version, which pushes the searcher either into the AI’s suggested business directly or toward a paid ad instead of an organic listing.

That’s a meaningful difference from a normal ranking drop. If you fall from position one to position three in the traditional pack, you’re still visible, still clickable, still in the conversation. If an AI Overview replaces the pack on that query and picks a competitor, you may not appear at all, regardless of how you’d have ranked in the format Google used to show.

This is part of why call and click volume can shift even when a business’s underlying local pack position hasn’t moved. The surface itself changed under them.

What to actually do about it

None of this calls for a separate “AI SEO” strategy running alongside your local SEO. The signals are the same signals. What changes is the priority you put on the specific, freshness-driven parts of your profile rather than the static parts you fill in once and forget.

  • Fix your primary category first if there’s any ambiguity in it. It’s the single input both systems weight most heavily for relevance.
  • Fill out services and products with real specifics, not placeholder text.
  • Respond to reviews close to when they’re posted, not in a monthly batch. Recency is doing real work here.
  • Keep photos and posts current. Treat the profile as a living channel, not a one-time setup task.
  • Match your schema to your GBP data, if you run your own website. Inconsistent data creates ambiguity Google has to resolve some other way.
  • Track your AI-surface share in Search Console, filtering search appearance by type, so you can see how many impressions come from AI-driven surfaces versus traditional results. If that share is rising for your service category, and it likely will keep rising, that’s a signal to keep watching, not a one-time fix to make and forget.

The 3-pack isn’t dead. But it’s no longer the only door your customers walk through, and for a growing set of queries, it isn’t even the first one they see. Treating your GBP as raw material for two systems instead of one is the actual shift here, not some separate AI-optimization checklist bolted on top of what you were already supposed to be doing.


About the author: Arif Rahman is a Dhaka-based freelance SEO strategist who handles local SEO for service and trade businesses across Bangladesh, Australia, and the UAE. He tracks how Google Business Profiles feed AI-driven search surfaces as part of ongoing GBP work for his clients, rather than treating it as a separate trend to chase.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top