Ask an AI which brand to buy and the shape of the answer depends on what you sell. Across 631,087 AI brand recommendations Parse extracted from ChatGPT and Google AI Overviews, in consumer electronics AI names one brand and stops: the single most-recommended brand holds 81% of the recommendation weight on the typical question. In sales and marketing software, the same exercise returns a wall of 15 names with no clear leader: the top brand holds just 22%. AI is not equally decisive across markets, and that changes what winning even means.
Some categories get one answer, others get a list
We measured this across 631,087 brand recommendations that Parse extracted from ChatGPT and Google AI Overviews answers between October 19, 2025 and April 24, 2026: every instance where one of those models recommended a specific brand for a buyer need. We grouped them into 10,839 distinct buyer questions, assigned each to its industry, and for each question measured how much of the recommendation weight landed on its single most-recommended brand across all the times we ran it. Call that the top-brand share. A high share means AI converges on one name; a low share means it spreads across a crowd. The typical buyer question sits in the middle: AI recommends 11 different brands across repeated runs, and the leader holds 33% of the weight. But the average hides the real story, which is how far the two ends pull apart.
- The typical AI buyer question draws 11 distinct recommended brands, and the single most-recommended brand holds 33% of the recommendation weight.
- Consumer categories are winner-take-most: in consumer electronics the top brand holds 81% of recommendation weight and AI names a median of 3 brands per question.
- B2B software is a free-for-all: in sales and marketing the top brand holds 22% and AI names 15 brands per question; software, IT, data, and AI all cluster near a 25% leader share.
- The gap holds on both ChatGPT and Google AI Overviews independently, and survives when we restrict to questions where AI named at least three brands.
- In a concentrated category, the goal is to own the one slot. In a fragmented one, the goal is to make the list and grind share, where being "not number one" is the normal state.
How concentrated is AI in your industry?
The spread is wide and orderly. We ranked 38 industries with at least 30 buyer questions by their median top-brand share. Consumer and physical-goods categories sit at the top: AI hands back a decisive pick. Business software, IT, and marketing categories sit at the bottom: AI hands back a slate. The "% decisive" column counts questions where one brand holds at least half the weight; "% open" counts questions where the leader holds under 30%.
| Industry | Median top-brand share | Brands named per question | % decisive | % open |
|---|---|---|---|---|
| Consumer Electronics | 81% | 3 | 84% | 5% |
| Consumer Goods | 67% | 5 | 63% | 21% |
| Clothing and Apparel | 65% | 6 | 62% | 18% |
| Sports | 63% | 5 | 64% | 23% |
| Food and Beverage | 50% | 8 | 52% | 29% |
| Real Estate | 50% | 7 | 51% | 28% |
| Health Care | 44% | 8 | 47% | 33% |
| Media and Entertainment | 36% | 9 | 37% | 41% |
| Commerce and Shopping | 34% | 11 | 34% | 43% |
| Financial Services | 30% | 11 | 19% | 49% |
| Software | 28% | 13 | 15% | 55% |
| Artificial Intelligence | 27% | 13 | 18% | 57% |
| Professional Services | 26% | 13 | 18% | 60% |
| Information Technology | 26% | 13 | 14% | 61% |
| Data and Analytics | 25% | 13 | 16% | 62% |
| Advertising | 22% | 14 | 10% | 76% |
| Sales and Marketing | 22% | 15 | 12% | 73% |
Read top to bottom, the table is a gradient from "AI knows the answer" to "AI lists the field." Consumer electronics is nearly three times as concentrated as sales and marketing, and names a fifth as many brands. If you sell software, you live in the bottom third of this table whether you like it or not.
If you want to see how AI engines describe your own brand, run a free brand check — it takes a minute.
Why physical-goods categories get a decisive answer
The pattern tracks how buyers and models treat the categories. Consumer-product questions often have a culturally agreed answer (a dominant TV brand, a default robot vacuum, a category-defining apparel name), and AI mirrors that consensus by naming it and moving on. In consumer electronics, 84% of questions return a decisive pick and only 5% return an open field. Part of this is question style: some consumer queries name one specific product, which mechanically concentrates the answer. But the effect is not an artifact. When we restrict to questions where AI named at least three different brands (genuine slates, not single-product lookups), consumer electronics still hands the leader 67% of the weight, while sales and marketing gives it 22%. The decisiveness is real, and it means a consumer brand that owns its category in AI answers owns almost all of the recommendation surface. There is no comfortable middle of the pack.
Why B2B software is a free-for-all
Business software runs the opposite way. Software, information technology, data and analytics, professional services, and artificial intelligence all cluster near a 25% top-brand share with 13 brands named per question. Sales and marketing is the most fragmented category we measured: AI names a median of 15 brands and the leader holds 22%. The mechanism is the market itself. These categories are crowded with near-substitutes, review coverage is dense and contradictory, and AI hedges by listing many credible options instead of committing to one. The practical consequence is that in software, being recommended in only a quarter of answers can still make you the category leader. A brand reading its own numbers against a consumer-goods benchmark will conclude it is losing when it is actually winning. This is the same reason a single blended AI visibility score misleads: a 25% share means very different things in a winner-take-most market and a fragmented one.
The pattern holds on both ChatGPT and Google AI Overviews
A concentration finding is only useful if it is not an accident of one model or one sampling choice, so we checked it three ways. First, by platform: the consumer-versus-software gap appears on ChatGPT and Google AI Overviews independently. In consumer electronics, ChatGPT gave the top brand 88% of the weight and Google AI Overviews 73%; in sales and marketing both sat near 21-24%. ChatGPT is consistently a touch more decisive than Google AI Overviews in concentrated categories, and the two are nearly identical in fragmented ones. Second, by run count: industries differ in how many times we ran each question (a median of 7 to 14), so we re-measured on only the questions with 9 to 13 runs, and the gap barely moved (consumer electronics 74%, sales and marketing 20%). Third, by slate size, restricting to questions where AI named three or more brands, which removes single-product lookups. The ranking survived every cut. The shape of an AI answer is a stable property of the category, not noise.
What this means for your AI visibility strategy
Find your row in the table first, because it sets the game you are playing. If you sell in a concentrated category (consumer electronics, apparel, sporting goods, most physical products), there is one slot worth having and it holds most of the recommendation weight. Second place is close to invisible, so the work is becoming the default answer: own the comparison and "best of" sources AI leans on, and accept that incremental share gains are rare and decisive ones are everything. If you sell in a fragmented category (software, IT, marketing, data), stop chasing a number-one share that the category does not produce. The leader there holds a quarter, so the winning move is to reliably make the slate of brands AI names, then grind your share up within it. Track whether you are in the list at all before you track your rank in it. The white space of buyer questions no brand owns yet is widest in exactly these fragmented categories, which is where new entrants have the most room.
How Parse measures this for your category
Parse tracks AI visibility across ChatGPT and Google AI Overviews, covering a public index of more than 4.7 million AI responses, 603,000 brands, and 57 million citations. The table above is the cross-industry pattern; your category has its own concentration number, and it decides whether your visibility plan should aim for a slot or a slate. Parse's Rankings view shows how recommendation weight is distributed across the brands AI names for your buyer questions (whether one name dominates or a dozen split it), and the Brand Lookup shows where you land in that distribution. If your category is concentrated and you are not the leader, the gap to close is the comparison and review surface that feeds the default answer; if it is fragmented, the gap is simpler coverage across more of the questions. For the related question of how many names AI lists at all, see our breakdown of how many brands a typical AI answer names. See how concentrated AI recommendations are in your category.
How we measured it and what it does not cover
This study reads Parse's recommendation-evidence layer: structured records of each time ChatGPT or Google AI Overviews recommended a specific brand for a buyer need, extracted from the answer text. We deduplicated to one record per answer, brand, and need, leaving 631,087 recommendation instances across 10,839 questions and 49,261 brands, and analyzed the 9,658 questions with enough volume to judge. Top-brand share measures recommendation weight across repeated runs, so it captures two things at once: how few brands AI names and how consistently it names the same leader. Both are facets of decisiveness, which is the property we report.
A few honest limits. Brands are extracted from answer text, so the counts reflect what AI said, but the set of questions is Parse's monitored prompt panel, so the industry mix reflects what we track, not the whole economy. The window covers ChatGPT and Google AI Overviews; the newer ChatGPT Search and Google AI Mode surfaces are not yet in this recommendation layer at scale. Each question is assigned to one industry via its dominant niche, and all figures are aggregate and k-anonymous. The numbers describe structure over a six-month window, not a live snapshot of any single brand today.
Does AI recommend the same brands every time?
It depends on the category. In Parse's data, consumer-product questions return a decisive single brand: in consumer electronics the top brand holds 81% of recommendation weight across repeated runs. Business-software questions return a rotating slate: in sales and marketing the leader holds only 22% and AI names a median of 15 brands. The typical question across all industries draws 11 brands with a 33% leader.
How many brands does AI recommend per question?
The median buyer question draws 11 distinct recommended brands across repeated runs, but the range is wide. Consumer electronics questions average a median of 3 brands; sales and marketing, advertising, and events questions reach 14 to 15. More crowded, substitutable markets produce longer lists; categories with a clear consensus leader produce short ones.
Which industries are most concentrated in AI recommendations?
Consumer and physical-goods categories. Consumer electronics is the most concentrated (81% top-brand share), followed by consumer goods (67%), clothing and apparel (65%), and sports (63%). In these categories one brand captures most of the recommendation weight and second place is close to invisible.
Why is my AI visibility share low even though I rank well?
Likely because you sell in a fragmented category. In software, IT, data, and marketing, the most-recommended brand holds only about a quarter of the weight, so a 25% share can make you the category leader. Comparing that number against a consumer-goods benchmark, where leaders hold 60-80%, makes a winning position look like a losing one.
What should I do differently in a concentrated versus fragmented category?
In a concentrated category, aim for the single slot: own the comparison and best-of sources that feed the default answer, because second place earns little. In a fragmented category, aim to make the slate first and grow your share within it. Track whether AI names you at all before you track your rank.