Ranked in the Map Pack? Google May Not Even Show You
For fifteen years, one target ran the entire local SEO conversation: get into the 3-pack, the little block of three businesses Google shows under the map. Agencies sold it. Owners bought it. That block is now shrinking to one or two names on a growing slice of searches, and the businesses that survive the cut are not the ones with the most keywords crammed into their profile. What the numbers show: your pack ranking is no longer a promise that anyone sees you, and chasing it harder is fighting a war that already ended.
The map pack's finish line moved from three names to one. Nobody sent you the memo, so you are still sprinting toward the old one.
Google is quietly cutting the pack from three names to one
Sterling Sky's Joy Hawkins published "The State of Local SEO in 2026" on June 26 and updated it two days later, and the finding underneath it should stop any owner mid-scroll. AI-powered local packs now show up on roughly 7% of tracked keywords on mobile in the US. When one appears, it does not list three businesses. It lists one or two. It carries no call button. And it frequently surfaces different businesses than the traditional pack sitting below it.
Read that last part twice. The AI pack and the classic pack can name different companies. So you can hold a solid position in the 3-pack you have watched for years, open your phone, and find a single competitor named above you in an answer that never mentions your business at all. Your rank tracker still shows green. The customer never scrolled far enough to see it.
Seven percent sounds small. It is not staying at seven percent, and even at today's level it lands hardest on exactly the high-intent searches where a phone call turns into a booked job. A single AI answer that names one plumber instead of three is a two-thirds cut to the businesses shown, applied to the moment a customer is ready to act. You did not lose a ranking. You lost the shelf.
The math on how many businesses just got cut
Vague claims about "fewer results" are easy to shrug off, so look at the count. Sterling Sky cited a Places Scout analysis that pulled the businesses named across two formats. Traditional 3-packs featured 18,330 unique businesses. AI local packs featured 5,943. That is 32% as many companies getting named, which is another way of saying about two-thirds of the businesses that used to earn a slot no longer appear when the AI answer takes over.
This was not a fluke in one city. Across 322 markets in the same analysis, 88% showed fewer unique businesses in the AI local pack than in the traditional pack. Nine markets out of ten, fewer names on the shelf. If you run a business in a mid-sized metro and you have been treating your pack ranking as money in the bank, that 88% is the sound of the bank quietly changing the locks.
Put it in plain terms. For every three businesses the old pack was willing to name, the AI version names about one, and the rest get sent to a page nobody reaches. Being "in the running" under the old rules means nothing if the race got shorter and you were not one of the names it kept.
Ranking held, the phone went quiet, and that is the whole problem
The uncomfortable truth is that ranking and getting contacted stopped moving together. Hawkins put it directly:
"If you have a business profile that ranks really well, even if you don't see a drop in ranking, your calls from the business profile are going down over time."
Joy Hawkins, Owner of Sterling Sky and Google Business Profile Product Expert, in "The State of Local SEO in 2026," June 2026.
She is not speaking in theory. Sterling Sky, working with data from Jepto, tracked 179 Google Business Profiles across 34 US law firms over two years. Rankings held in plenty of those cases. Click-to-call dropped anyway, and stayed down. If a category as competitive and cash-rich as legal services cannot buy its way out of this with strong rankings, your local shop is not the exception.
Part of the reason is that Google has been physically removing the call button. Sterling Sky documented handyman profiles losing their call button in favor of images between October and November 2025, following the same pattern dentists saw earlier. A customer looking at a handyman used to tap a green call button. Now they see a photo. The path from "found you" to "phoned you" got a step longer, and every added step leaks customers.
I have written before about the gap between a healthy rank report and a silent phone, because the disconnect between your ranking and your ringing phone is now the most common thing owners get blindsided by. The AI pack pours accelerant on it.
What actually earns the one slot AI is willing to name
When an answer engine has room for a single business, it does not roll dice. It names the business it is most confident about, and confidence is built from corroboration. Darren Shaw of Whitespark described where AI discovery pulls from bluntly: it draws on "what the entire internet says about you." Not what your profile claims. What the web confirms.
That is the pivot most owners miss. A profile stuffed with service keywords is a business talking about itself. Reviews, mentions on other sites, consistent name-address-phone data, and a real-world footprint are the internet talking about you, and the AI trusts the second one. Hawkins did not soften her read on where this is heading. On the new AI local packs, she said plainly:
"These things are terrifying."
Joy Hawkins, Sterling Sky, on AI-powered local packs, June 2026.
The gap is wider than most owners assume. Danny Goodwin reported in Search Engine Land on January 28, 2026 that AI assistants recommend only 1% to 11% of the businesses that rank in traditional Google results, making it up to 30 times harder to get surfaced by AI than to rank the normal way. Thirty times. That is not a tweak to your checklist. That is the checklist being rewritten while you were still filling out the old one.
So this is what actually moves the needle, in order of what pays back fastest.
First, make reviews a weekly habit, not a launch event. A steady drip of recent, specific reviews signals a live, real business to both the classic pack and the AI answer. Ask every satisfied customer the day the job closes. If you delegate this, the instruction to your team is one sentence: request a review from every completed customer within 24 hours, by text, with a direct link. I have made the case that reviews now beat proximity for winning the local top three, and the AI pack raises the stakes on that further.
Second, fix your name, address, and phone number everywhere they appear online, and make them identical down to the punctuation. Your website, your profile, your directory listings, your social pages. Inconsistency reads as uncertainty to a machine deciding whether to trust you enough to name you alone. Pull up your top ten listings and compare them line by line, or hand that exact task to whoever manages your marketing.
Third, get mentioned by other local sites you do not own. A local news write-up, a supplier's partner page, a chamber directory, a sponsored little-league page. Each independent mention is a vote the AI can count. Spend an hour listing every local organization you already have a real relationship with, then ask each one to link to or name your business.
Fourth, answer the real questions customers ask, in public, where the AI can read them. Fill out your services in full, keep your hours accurate, and use your profile's Q&A and posts to state plainly what you do and where. If you are handing this to an agency, ask them one question: "What specifically are you doing to get us named in AI-generated local answers, not just ranked in the pack?" If the answer is a blank stare or a rankings screenshot, you have found the problem.
Fifth, decide carefully how much of your review voice you automate. Google's own AI can now draft your review replies, and there is a real trade between speed and sounding like a human who was actually there. Automate the whole thing and every reply starts to read the same, which is exactly the kind of hollow signal AI is learning to discount. Keep a human editing the ones that matter.
What to measure now, and the number that will lie to you
Rank tracking is the vanity metric of this era, and it is about to fool owners into thinking nothing is wrong. Your position can sit at number two in the classic pack while an AI answer above it names one competitor and never shows you. A green rank report and a dead phone can coexist now, and if you only watch rank, you will not see the loss until the quarter's revenue does.
Watch contact actions instead. In your Google Business Profile insights, track calls, direction requests, and website clicks month over month, and treat a sustained decline as the alarm even if your ranking is flat. That is the exact pattern Sterling Sky found in the law firm data: rank steady, calls sliding. If you see the same shape in your own numbers, believe the calls, not the rank.
Set a simple monthly rhythm. First of the month, pull last month's calls, direction requests, and clicks and write them in one place. Run your top three service searches on your own phone, on mobile data, not on office wifi that skews location, and note whether an AI pack appears and whether you are in it. Count your new reviews for the month. Three numbers, one search, ten minutes. That beats any dashboard an agency will charge you to stare at.
One trap to name directly. A high star average is not the same as a live review profile. Recent reviews carry weight that a frozen five-star history does not, which is why a moving 4.6 now beats a stale 4.9. When you measure reviews, measure how recent the last one is, not just the average across all time.
Frequently Asked Questions
If I still rank in the 3-pack, am I safe from the AI local pack?
No, and that assumption is the trap. Sterling Sky's 2026 analysis found that AI local packs frequently surface different businesses than the traditional pack, so you can hold a strong 3-pack position and still be absent from the AI answer sitting above it. The two are scored differently. A customer who sees the AI answer first never scrolls to the pack where you rank well.
How common are these AI local packs right now?
As of Sterling Sky's June 2026 report, AI-powered local packs appeared on roughly 7% of tracked keywords on mobile in the US. That is a minority of searches today, but it concentrates on high-intent queries and shows only one or two businesses with no call button. Treat 7% as a floor to prepare for, not a ceiling to relax under.
What one thing should I do first if AI local answers scare me?
Start a weekly review habit. Darren Shaw of Whitespark noted that AI discovery pulls from what the entire internet says about you, and steady recent reviews are the strongest corroboration you control directly. Ask every completed customer for a review within 24 hours by text with a direct link. It is the single fix that helps both your classic pack ranking and your odds of being the one business AI names.
The old scoreboard said you were winning. It measured a race that is quietly being replaced by one with a single lane. The businesses that will still get phones ringing a year from now are not the ones defending a pack position with more keywords; they are the ones the whole web already agrees on. Which raises the question worth sitting with: if a machine had to name exactly one business in your category tomorrow, and it only trusted what the internet already says about you, would it say your name?
