AI local packs name fewer businesses

Google has been rolling out AI-powered local packs on mobile in the US. They often name one or two businesses instead of three, sometimes without a call button. Sterling Sky’s Places Scout analysis found 5,943 unique businesses in AI packs versus 18,330 in classic three-packs on the same reports, about 32% as many.

Sterling Sky published the count. Places Scout looked at ranking reports and found 5,943 unique businesses in AI local packs against 18,330 in the classic three-pack. Roughly a third as many names get shown. In 88% of 322 markets, the AI pack was the narrower list.

I read the actual announcements instead of the coverage.

These packs are not replacing every map result. On the keywords they tracked, AI local packs showed up on about 7%, mobile, US. The classic pack still owns a lot of ready-to-call “near me” intent. The AI format is a shorter list that often picks different businesses, and often drops the call button.

Classic three vs the short list

The map pack used to show three names with room to call. The AI version often shows one or two. When the seat count drops, the businesses that stay visible are the ones the system can trust from profile, reviews, and site that agree.

Most people don't have a workload problem. They have a systems problem.

What the count actually measured

The Sterling Sky write-up is researcher data from Places Scout ranking reports, not a Google official share of all queries. Treat the numbers as what that tracked set showed: fewer unique businesses in the AI packs, and a different set of winners in many markets. Cite the Sterling Sky state of local SEO note when you use the figures. Do not stretch them into a universal percent for every trade and every city.

Whitespark’s intent split, cited in 2026 local coverage of 540 queries, still puts the Map Pack on top for ready-to-call local intent. AI Overviews cover more of the informational and hybrid research queries. Parallel systems. Same underlying material.

Presence that both systems can use

Every business owner is the rate-limit on their own business. The way out isn't another hire.

The rate limit here is presence both systems can use. Profile, reviews, and site that match. A rule without a script is a wish. “Be findable” without a listing and a site that match is a wish. The handover is not ground truth. A rank report that looked fine last month does not prove the AI answer will name you this week.

Confidence is not correctness. Ranking in the old pack does not mean the short list names you.

What to do next

If the basics are broken, fix them before you buy more noise. Get Seen is the Gman Creations path for website and Google-presence foundations, including search and AI-readable content basics. If you want a quick read on homepage basics search and AI tools can parse, SignalGrade is the free check. Neither invents a seat in the pack. Both start from what is actually on the page and the profile.

When the short list only has room for one or two names in your trade, are you on it, or are you just assuming you are?

FAQ

Are AI local packs replacing the map pack?

Not on every query. In the Places Scout set Sterling Sky reviewed, AI local packs showed on about 7% of tracked keywords on mobile in the US. The classic pack still owns a lot of ready-to-call near-me intent.

Does ranking in the three-pack mean the AI pack will name you?

No. The AI format often picks different businesses, and the list is shorter. A last-month rank report is not proof the short list names you this week.

What should a local business fix first?

Make the Business Profile, reviews, and site agree on the same facts. A thin profile and a vague site leave both the classic pack and the AI answer with little to trust.