AI Overviews by Industry: Which Sectors See the Highest Citation Rates

Which industries appear most in Google AI Overviews?

Health, people and society, and science appeared most often in Semrush's desktop category analysis of AI Overview results. Semrush studied 100,000 US desktop keywords and 100,000 US mobile keywords, then reviewed category-level AI Overview presence in its sample. On desktop, health ranked first, people and society second, and science third. On mobile, people and society ranked first, science second, and food and drink third.

That result is useful as a directional map, not a universal market share report. Semrush filtered its keyword set to queries that had triggered an AI Overview at least once between September 1 and 10, 2024. It therefore describes the mix within that research sample. It does not establish that the same category order applies to every keyword set, country, device, or later Google result page.

For teams working on GEO, the practical point is simple: category labels are less useful than the questions inside them. A health publisher may have many questions that Google treats as explanatory. A software company may have a few narrow comparison questions that trigger regularly. Measure the question set you need to own before treating an industry ranking as a forecast.

1. Desktop leaders in Semrush's category analysis: health, people and society, science.
2. Mobile leaders in the same analysis: people and society, science, food and drink.
3. The source examined US keywords that had already triggered at least one AI Overview during its September 2024 collection window.

Do health or technology queries trigger AI Overviews more often?

Health queries appeared more often than technology-adjacent categories in Semrush's desktop sample. Health ranked first, while internet and telecom ranked fourth and computers and electronics ranked fifth. The result supports a cautious conclusion about that dataset: health had stronger observed AI Overview presence than those two technology categories on desktop.

Technology is not absent from the picture. Internet and telecom, along with computers and electronics, were both within the top seven mobile categories in the Semrush chart. That means a technology team should not read health's desktop lead as a reason to ignore AI Overviews. It should separate informational support queries from product, login, pricing, and branded queries, then inspect each set in live results.

Google does not publish a sector rule that guarantees an AI Overview for health or technology. Google says AI Overviews appear only when its systems determine they add something beyond classic Search, and notes that they often do not trigger. That is why AEO starts with answering the underlying question clearly instead of treating an industry label as a ranking factor.

1. Health ranked first on desktop in Semrush's category chart.
2. Internet and telecom ranked fourth on desktop and sixth on mobile.
3. Computers and electronics ranked fifth on desktop and seventh on mobile.

What does the public data mean by an AI Overview trigger rate?

A trigger rate should mean the share of a defined query set that produces an AI Overview during a specified observation window. That definition needs the keyword universe, country, device, date, and method. Without those fields, a percentage can look precise while answering no reproducible question.

Semrush's published category chart uses desktop and mobile weight. The article explains that it reviewed Semrush Sensor keywords by category and weighted categories against each other to estimate which industries most often triggered AI Overviews. It does not publish an exact industry-by-industry trigger percentage in the article text or chart labels. The visual shows relative weight, not a reusable percentage denominator for every category.

Clear measurement boundaries help answer engines teams interpret results. A category can lead a sampled chart because its questions occur frequently among observed AI Overview keywords. That is different from proving that a randomly selected query in that category has a given probability of producing an overview. Keep presence, rate, and citation outcomes separate in reporting.

1. Use a query-set denominator for a true trigger rate.
2. Record device, market, dates, query intent, and collection method.
3. Do not relabel a category-weight chart as a citation rate study.

What is citation diversity, and does Google publish it by industry?

Citation diversity measures how widely supporting links are distributed across distinct domains, publishers, or source types in a defined set of AI answers. A useful index needs a declared unit. It may count unique domains per answer, calculate concentration across all answers, or compare the share held by the most-cited domains. Each method answers a different question.

Google says AI Overviews can surface relevant links and may use query fan-out to identify supporting web pages. Google also says AI has can surface a greater variety of links than classic Search. Those statements explain why source diversity is worth measuring. They do not provide a public, industry-level citation diversity index.

No public source in the cited evidence publishes a 12-industry AI Overview citation diversity index. The table below marks that field as not published rather than assigning invented scores. A defensible Citation Share study would collect a defined prompt set, preserve every supporting URL shown in each answer, normalize domains, and state its diversity formula before ranking sectors.

1. Citation diversity is a measurement design, not a Google-published industry metric.
2. Distinct domains and total links are different measurements.
3. An index requires a disclosed query set and a disclosed formula.

Which 12 industries have the clearest published AI Overview presence signal?

The 12 categories below are the first 12 desktop categories in Semrush's published chart, ordered by the chart's desktop weight. They are not ranked by opportunity score because the source does not publish an opportunity-score calculation. They also are not ranked by citation rate because it does not publish citation rates.

This is still a useful audit table. It shows which sectors have a public presence signal and where the evidence stops. A team can use it to avoid turning a chart into a stronger claim than the source permits. The highest desktop category is health. The lowest category among these selected 12 is travel, but that does not make travel the lowest category in Semrush's complete desktop chart.

A real opportunity score can be built for one business, but it should combine its own answer presence, Citation Share, commercial relevance, and the difficulty of producing a better source. Those inputs belong to a monitored query universe. They cannot be inferred from an industry chart alone.

1. The table preserves Semrush's published desktop category order.
2. Exact category trigger percentages were not published in the source.
3. Citation diversity and opportunity score are deliberately left unscored.

What does the 12-industry evidence table show?

The evidence table shows a visible presence ranking, not a citation-rate ranking. Health, people and society, and science sit at the top of Semrush's desktop chart. Internet and telecom and computers and electronics follow, which places two technology-adjacent categories near the front of the observed desktop distribution.

Every row has the same evidence limitation. Semrush's chart does not show the exact percentage for each industry, and neither Semrush nor Google publishes a citation diversity index for these categories. Filling those cells with estimates would turn an editorial study into invented data.

The mobile column is included as a second observed rank where the category appears in the mobile chart. It makes the device difference visible. Food and drink, for example, ranked third on mobile but sits outside this desktop top-12 selection. That is a reason to keep device-specific reporting rather than collapse desktop and mobile into one score.

1. Read desktop rank as source-observed category weight rank.
2. Read mobile rank as a separate source-observed category weight rank.
3. Treat unavailable figures as unavailable, not as zero.

What sector has the lowest AI Overview citation rate?

No cited public study identifies a sector with the lowest AI Overview citation rate. The Semrush research measures category weight among a keyword sample with observed AI Overviews. It does not measure how often each industry is cited, how many domains are cited, or the share of all category queries that generate an overview.

For the narrower issue of which category sits lowest in Semrush's published charts, the answer varies by device and is not a citation finding. In the desktop chart, hobbies and leisure is the lowest category with a visible bar. In the mobile chart, shopping is the lowest category with a visible bar. Those are source-chart observations, not citation rates or broad claims about business opportunity.

Do not make a low-observed-presence category disappear from a monitoring program. Google says AI Overviews only show when its systems judge them additive to classic Search. A sector may have fewer broad explanatory queries and still have high-stakes comparison or local questions where an overview changes the path to a buyer.

1. No source cited here publishes a lowest industry citation rate.
2. Desktop lowest visible category in the Semrush chart: hobbies and leisure.
3. Mobile lowest visible category in the Semrush chart: shopping.

Why do informational questions matter more than industry labels?

Informational query shape is the stronger public signal in Semrush's dataset. The study found that 80% of its desktop AI Overview keywords and 76% of its mobile AI Overview keywords had informational intent. It also found that 35% of desktop keywords and 32% of mobile keywords were question keywords.

This does not mean every how-to query earns an overview. It does mean that an editorial program should map the actual questions people ask before allocating content by industry label. Start with explanatory questions, process questions, technical definitions, and comparison questions that require several supporting facts. Then check live results over time.

Google's guidance reinforces the restraint. Google says there are no additional technical requirements for eligibility beyond being indexed and eligible to appear in Google Search with a snippet. Existing SEO best practices remain relevant. Strong source material, clear pages, internal links, and technical eligibility are concrete work. There is no special markup that guarantees a citation.

1. Semrush: 80% informational intent on desktop.
2. Semrush: 76% informational intent on mobile.
3. Google: no additional technical requirements exist for AI Overview eligibility.

How should a team measure AI Overview opportunity by sector?

Measure AI Overview opportunity with a fixed query panel for the business, not a generic sector score. Split the panel by audience, intent, device, market, and question type. Preserve the result page date because Google can vary the answer and links over time. That gives the team a repeatable baseline instead of a one-off screenshot.

For each query, record whether an overview appeared, which domains received supporting links, whether your domain appeared, and whether a competitor appeared. Calculate Answer Presence for the share of queries where an overview appeared. Calculate Citation Share for the share of relevant AI answers that cite your domain. Review the source-domain distribution separately rather than calling link count diversity.

The result is more useful than an industry chart because it connects the measurement to a decision. A B2B software company can track category, alternatives, implementation, and integration questions. A local service business can track service-and-location questions. Each group can then see where coverage or evidence is missing and decide what to publish next.

1. Define the query panel before collecting results.
2. Store source URLs and domains with every observed answer.
3. Report Answer Presence, Citation Share, and source diversity as separate metrics.

What should publishers do with the industry findings now?

Publishers should use the industry findings to prioritize research, not to claim a guaranteed route into Google AI Overviews. Health's desktop lead and technology's upper-tier positions show that explanatory demand can produce frequent AI Overview exposure in a sampled dataset. They do not replace a live audit of the questions that matter to a specific site.

Build pages that answer one question directly, show the supporting evidence, state limits honestly, and connect related pages with useful internal links. Google says that the same foundational SEO practices remain relevant for AI features. That is a more durable operating rule than any claim about an industry having a permanent citation advantage.

The next measurement step is concrete: create a sector-specific query panel, run it on a fixed schedule, and log the overview appearance plus supporting domains. That turns a public category chart into an evidence base for editorial choices without overstating what the chart proves.

1. Use public research to frame hypotheses.
2. Validate hypotheses against a defined live query panel.
3. Keep citation, presence, and traffic reporting distinct.

Sources

Source: Semrush

Semrush studied 200,000 US keywords, reported industry category rankings, and found informational intent in 80% of desktop and 76% of mobile AI Overview keywords. Semrush, 2025-07-22

Source: Google Search Central

Google says AI Overviews only appear when its systems determine they add value beyond classic Search, and that no additional technical requirements exist for eligible pages. Google Search Central, 2025-12-10

Source: Pew Research Center

Pew Research Center found that Google users clicked a traditional search-result link in 8% of visits with an AI summary and 15% of visits without one in its March 2025 browsing analysis. Pew Research Center, 2025-07-22