A 1,120-Query Study Found a 28.5% Local Pack-to-AI Mode Visibility Gap

A healthy Google Maps position does not prove that the same business appears in Google AI Mode.

In one 2026 vendor study, researchers ran 1,120 paired query executions across seven service verticals and 13 U.S. cities. They found that 28.5% of businesses returned in the local pack were absent from AI Mode for the same sampled query.

That is a result from a defined study window, not a universal percentage for every business or market. Its practical value is still significant: a local pack report and an AI Mode check measure different visibility outcomes.

Start with the complete Google Maps visibility diagnostic, then add AI Mode to the measurement set instead of assuming one surface represents the other.

The direct answer

Does a local pack ranking guarantee AI Mode visibility? No.

The Local SEO Data study collected its 1,120 query executions from May 20–22, 2026. It compared AI Mode business results with traditional local pack results across plumbers, dentists, HVAC, chiropractors, lawyers, restaurants, and medical specialists.

The researchers reported:

Finding	Study result	What it means
Local pack businesses also present in AI Mode	71.5%	Most overlapped, but a substantial share did not
Local pack-to-AI Mode gap	28.5%	More than one in four pack businesses in the sample was absent from AI Mode for the paired query
Top-three result-set overlap	48%	The two top-three sets differed materially; this is not a same-rank retention rate
Businesses found only in AI Mode	999	AI Mode also surfaced businesses outside the sampled local pack results
Median businesses shown	5 in AI Mode versus 3 in the local pack	The compared result sets had different sizes

The study disclosed that it used Local SEO Data's own API and published methodology, limitations, confidence intervals, and raw-data access information. That transparency helps readers evaluate the result, but it remains vendor research rather than a Google explanation of its internal systems.

Do not misread the 48% number

The 48% figure is top-three set overlap.

It does not mean that a business keeps the same rank less than half the time. It does not measure position retention at all. It says that when the researchers compared the top three businesses from each surface, 48% of the two sets overlapped.

That is still important. A business can be prominent in one result set and absent from the other. But the correct conclusion is result-set divergence, not a ranking-loss probability.

AI Mode, AI local packs, and AI Overviews are not the same thing

Local visibility reporting gets confusing because several Google surfaces now use “AI” language.

Surface	What the cited research measured
Traditional local pack	The familiar map-oriented local business result set
AI Mode	Google's conversational search experience, which can return business results and supporting links
AI local pack	A compact business module observed inside AI-generated results
AI Overview	A generated summary that may include local businesses or third-party citations

A separate Sterling Sky State of Local SEO in 2026 study tracked U.S. mobile keywords and reported that AI local packs appeared for roughly 7% of those keywords. Across 322 tracked markets, the study counted 5,943 unique businesses in AI local packs and 18,330 in traditional packs—about 68% fewer—with 88% of markets showing a reduction.

Those figures describe Sterling Sky's tracked keyword and market set. They do not prove that every AI local module has the same size or that Google uses one fixed selection threshold.

AI Overviews change with query intent

Whitespark ran a different 540-query study of AI Overviews and local search in May 2025.

The sample covered Houston, Phoenix, and Denver across plumbers, personal injury lawyers, dentists, optometrists, medical clinics, and real estate agents. Within that sample:

Query class	AI Overview appeared	Local pack appeared
Clearly local intent	15%	93%
Informational intent	92%	6%
Hybrid intent	97%	17%
All sampled queries	68%	39%

The result does not show that “most local queries” always receive AI Overviews. It shows that feature prevalence varied sharply by the researchers' intent classification in one defined sample.

Whitespark also reported no apparent overall correlation between local pack prominence and AI Overview prominence in that sample. In one small nine-query Houston plumber subset, 60% of 23 citations went to third-party publishers. That subset is useful as an example of citation diversity, not a universal local-search rate.

Query phrasing can change whether listings appear

The Local SEO Data study included a useful phrase test.

Among 38 cases where a bare keyword produced no business listings, adding “near me” produced listings in all 38. That does not mean every “near me” query rescues every business. It means phrasing changed the presence of business listings in this specific test set.

Use multiple formulations for each priority service:

• Bare service: “water heater repair”
• Local modifier: “water heater repair near me”
• City modifier: “water heater repair Denver”
• Need statement: “who repairs a leaking water heater near me”
• Comparison or qualification: “best emergency plumber for water heater repair”

Visibility for one phrasing is not evidence of visibility across the others.

Review counts were associated, not proven causal

The Local SEO Data study also compared review counts.

In its consumer-service sample, businesses found only in AI Mode had 49% more median reviews than businesses found only in the local pack. For plumbers, the reported medians were 546 versus 125, described as a 337% difference.

That is an association. It does not prove that adding reviews caused AI Mode inclusion.

Businesses with more reviews may also have stronger brands, older profiles, more customers, wider citation footprints, better websites, or other advantages the comparison did not isolate. Professional-service verticals in the study showed a different pattern, another reason not to turn the correlation into a rule.

Use review evidence as one diagnostic input. Do not promise an AI Mode ranking from a review target.

Google does not publish an AI Mode local-ranking formula

Google's local ranking guidance describes local results through relevance, distance, and prominence.

Google's AI features guidance says the same foundational SEO practices remain relevant to AI Overviews and AI Mode. It also says there are no extra technical requirements or special markup required to be eligible.

Those pages do not disclose the weights or exact local-business selection logic used inside AI Mode.

It is fair to say the measured surfaces returned different businesses. It is not fair to claim that they use identical inputs, a particular undisclosed evidence layer, or known separate algorithms.

Consumer behavior makes the second measurement worth adding

BrightLocal's 2026 AI trust and local consumer review survey surveyed 1,002 U.S. adults.

In that survey:

• 45% of respondents reported using AI tools to find a local business, compared with 6% in the prior annual survey.
• AI ranked third behind Google and Facebook among reported tools used for local recommendations.
• Reported Google use was 71%, compared with 83% in the prior survey.

These were annual survey estimates, not necessarily the same respondents tracked over time. They indicate a large directional change worth monitoring; they do not establish the exact population-wide rate of change.

A measurement protocol that produces useful evidence

Do not replace one vanity number with another. Use a paired test.

1. Define the commercial query set

Choose 10 to 20 queries tied to real services, not only broad educational topics. Record the buyer intent and the page or offer that should satisfy it.

2. Define the geography

Record the city, neighborhood, or coordinates used for the local pack measurement. Note whether AI Mode respected or inferred the same location.

3. Run paired checks

For each query, record:

• Local pack businesses and positions
• AI Mode business mentions
• Owned-site links
• Third-party sources cited
• Whether the business is mentioned without a link
• Whether the answer contains inaccurate business facts
• Screenshot or saved answer evidence
• Date, device, account state, and query wording

4. Separate exact outcomes

Use separate fields:

Field	Example state
Local pack	Present at position 3
AI Mode mention	Absent
Owned citation	Absent
Third-party citation	Industry directory present
Business facts	Correct hours, wrong service area
Conversion route	Website link present; no call action

Do not flatten those into one score until the underlying rows remain inspectable.

5. Retest the same way

AI answers and local results change. A single snapshot is evidence of one moment, not a stable rank. Repeat the same query, geography, and documentation method after a defined window.

What to do when the surfaces disagree

If the business appears in the local pack but not AI Mode:

• Check whether the relevant website page is crawlable and useful.
• Compare the sources AI Mode cites for competitors.
• Correct conflicting business facts across major profiles.
• Strengthen direct answers, proof, and entity clarity on the relevant page.
• Confirm that reviews, services, categories, and location data are accurate.
• Retest the same prompt rather than switching to an easier one.

If the business appears in AI Mode but not the local pack:

• Audit verification, category, address eligibility, service area, and proximity.
• Check the local landing page and internal links.
• Measure a geographic grid rather than one search location.
• Treat the AI mention as useful visibility, not proof that the Maps problem is solved.

For the wider model, read SEO, AEO, and GEO for local businesses. For score boundaries, read what an AI visibility score can and cannot measure.

What this evidence can and cannot prove

The studies support a careful conclusion:

• A local pack position does not guarantee an AI Mode appearance.
• AI Mode can surface businesses outside the paired local pack set.
• Query phrasing and intent can change the visible surface.
• Review counts were associated with some study outcomes.
• Reporting should keep local pack, AI Mode, AI Overview, citations, and conversion paths separate.

They do not prove:

• A universal 28.5% gap for every industry
• Why Google's systems selected each business
• That more reviews cause AI Mode inclusion
• That an AI mention caused a lead
• That one successful query represents market-wide visibility

Next step

If you have never checked what AI systems say about the business, start with the free AI visibility score.

If you want local pack and AI surfaces measured with separate proof, see New Reward's monthly proof reporting.

The goal is not to collect more rankings. It is to know which buyer surfaces include the business, which evidence they use, and which approved fix should ship next.