Something Inc.Schedule a free consultation
STRATEGY

Some AI categories are already closed. Check yours.

In news and media, three brands hold 82.9% of AI visibility. In finance it is 41.4%. That single number should decide whether you play offense or pick a narrower fight.

TTTyler TruffiManaging Partner · AUG 13, 2026 · 10 MIN READ

I have had the same conversation four times this month. A team has done the work, shipped the comparison pages, fixed the crawl issues, and their AI visibility has barely moved. The instinct is to assume the execution was wrong. Sometimes it is. Often the category was simply further along than anyone checked.

Semrush's expanded 2026 AI Visibility Index gives you a way to check. They analyzed 126 million US AI search prompts from January through April 2026 across ChatGPT, Gemini, Google AI Mode, and AI Overviews, benchmarked across 22 industries. The finding worth stealing is not any single brand's score. It is how differently AI search visibility by industry concentrates depending on which industry you happen to be in.

82.9%
of news and media visibility held by three brands
41.4%
the same top-three figure in finance
36
brands visible on every platform, every month
THE SHORT VERSIONTop-three concentration is the single best predictor of whether a GEO program can move your numbers this year. Above roughly 75%, you are fighting for scraps and should narrow. Around 40%, the category is genuinely open and worth committing to.

The number that tells you if your category is winnable

Concentration is a familiar idea in market analysis and a strangely absent one in AI visibility work. Most programs measure their own mention rate against last month and call that progress. That tells you whether you improved. It does not tell you whether improving is going to matter.

The useful question is what share of all visibility in your category the top three brands already hold. If the answer is 83%, the remaining field is splitting 17% and your realistic ceiling is a slice of that. If the answer is 41%, more than half the category is unclaimed and the ceiling is a genuine number.

INDUSTRYTOP-THREE SHARE OF CATEGORY VISIBILITYWHAT THAT IMPLIES
News and media82.9%Effectively decided; compete on niche or syndication
Consumer electronics76.9%Crowded; win a subcategory before the category
Industrial42.2%Open; specificity beats brand size
Finance41.4%Open; the most winnable of the four

Sit with the finance number for a second, because it is counterintuitive. Finance is not a small or unsophisticated market. It has enormous incumbents with enormous budgets. And its top three still hold under half of AI visibility, which means engines are distributing answers across a wide field of sources rather than defaulting to the biggest names. That is what an open category looks like from the inside.

You are not losing to the leaders. In half of these categories, the leaders have not won yet.

AI search visibility by industry, ranked

Here is the same data as a picture, because the spread is the point and a table flattens it.

News and media83%
Consumer electronics77%
Industrial42%
Finance41%

Share of category AI visibility held by the top three brands (Semrush 2026 AI Visibility Index)

A forty-point spread between the tightest and loosest categories is not noise. It is two different games being played under one label, and generic GEO advice addresses neither of them well. The advice that works in finance, publish deeply, be specific, out-detail the incumbents, is close to useless in news and media, where three brands have already absorbed the answer space and the only remaining moves are narrower.

There is a second number in the index that makes the concentration read even sharper: only 36 brands maintained visibility across every platform for every month of the study. Thirty-six. Across 22 industries. Consistent cross-platform presence is not a table-stakes achievement that everyone has and you are missing. It is close to nonexistent, which is oddly good news, because it means the bar for durable multi-engine visibility is lower than the noise around it suggests. We looked at the same dynamic from the platform side in our study of source concentration across AI platforms, and the pattern held: consistency is rarer than dominance.

Classic search results have ten slots. AI answers have three to fifteen sources, and in Gemini's case an average of three. Fewer slots, harder cutoff.

1The slot count collapsedA page ranking eighth in Google still gets seen. There is no eighth position in a three-source answer. The distribution of attention went from a gentle curve to a cliff, and cliffs concentrate.
2Models reward corroboration, which compoundsEngines prefer sources that other trusted sources agree with. Brands already cited get cited more, get referenced more, and become easier to corroborate. That is a feedback loop, and feedback loops produce winner-take-most outcomes quickly.
3Categories with clear entities settle firstNews has obvious canonical publishers. Consumer electronics has obvious canonical brands. Industrial and finance are fragmented, specification-heavy, and full of narrow questions no single incumbent owns. Fragmentation is your friend here.

The retail data underlines how fast all this is moving. Adobe figures cited in the same index put retail AI traffic up 1,324% between October 2024 and May 2026, with travel up 2,215%. Categories are being sorted while the traffic is still multiplying, which is exactly when position is cheapest to take and most expensive to lose. We wrote about the retail end of that shift in where AI search visibility is moving toward retailers.

What AI search visibility by industry means for your plan

Run the concentration number for your category first, then pick the matching strategy. These are genuinely different plans, not intensities of the same plan.

Play to win
Open category, roughly 40% concentrationCommit. Build the full comparison set, the technical foundation, and the corroboration campaign at once. This is the rare situation where breadth pays, because no incumbent has claimed the answer space and depth compounds against an unfinished field.
Play to place
Crowded category, 75% and upNarrow before you build. Pick the subcategory, segment, or use case where the top three are generic, and own that completely. Category-level prompts are a losing fight; the specific prompt underneath them is often unclaimed.
Prerequisite
Any category, if you are not measuring by engineFix that before anything else. A blended score across engines that behave differently will make an open category look closed and a closed one look promising. Segment first, decide second.
Structural
Any category, if you run SEO and AI separatelyMerge them. In the index, organizations running fully integrated search and AI strategies reported increased traffic or leads 81% of the time, against 36% for those managing the two separately. That gap is larger than most tactical wins available anywhere else on this list.

That last card deserves more than a card. An 81% to 36% split between integrated and separated programs is the kind of finding that should reorganize a team, not just inform a tactic. It is also intuitive once you accept that citations and rankings draw on overlapping signals: the technical work that makes you crawlable serves both, the content that earns links earns corroboration, and splitting the two into separate teams with separate targets mostly produces duplicated effort and contradictory priorities. This is the argument we make when content marketing and technical work land in one plan rather than two.

The measurement gap underneath all of it

One more number from the index, and it reframes everything above: 45% of marketing leaders cannot accurately measure their brand visibility in AI answers, and only 9% have tools covering every relevant metric across platforms.

45%
of leaders cannot measure their AI visibility accurately
9%
have tooling covering every relevant metric
81% vs 36%
traffic gains, integrated vs separated strategies

So when I say check your category's concentration, I am aware that most teams currently cannot. That is the actual starting point for the majority of programs, and pretending otherwise produces strategy documents built on vibes. The work is not glamorous: assemble a frozen prompt set that reflects real buyer questions, run it across engines on a schedule, and record who appears alongside you. Your competitors' share is as informative as your own, and almost nobody logs it.

The 9% figure is the one I would put in front of a board. It says that in a channel everyone agrees is reshaping demand, nine out of ten organizations are flying without instruments. That is not a tooling problem so much as a sequencing mistake: teams commissioned content before they built the ability to tell whether content was working. The order should be reversed, and it is cheap to reverse. A frozen prompt set and a spreadsheet beats a dashboard you cannot interpret, and it can be running this week. The full index and its methodology are published in Semrush's announcement of the 2026 AI Visibility Index, and it is worth reading the industry cuts directly rather than trusting a summary of them, including this one.

There is a version of this argument that goes too far, and I want to name it before someone runs with it. Concentration is not destiny. A category at 80% top-three share is hard, not sealed, and the brands holding that share are frequently holding it on generic category prompts while losing every specific one underneath. The mistake is not entering a concentrated category. The mistake is entering it with a plan built for an open one, discovering after three quarters that the headline prompts were never available, and concluding that the whole channel does not work. It works. The target was wrong.

A caution on the figures themselves. Concentration numbers depend heavily on how the category was defined and which prompts represented it. Semrush's 22-industry cut is coarser than your actual competitive set, and your real category is probably narrower than theirs. Use their numbers to calibrate expectations, not to conclude. The version that should drive budget is the one you measure on your own prompts, where a category that looks closed at industry level frequently turns out to be wide open two levels down. That has been the pattern in most technical categories we have built into, including a security category build where the industry-level picture was far more crowded than the buyer-level one.

The honest next move

Take thirty prompts your buyers genuinely ask. Run them across ChatGPT and Gemini. Log every brand that appears, not just yours. Add up the top three brands' share of total appearances. That number is your category's concentration, measured on the only prompt set that describes your business.

Then be honest about what it says. If three names are taking most of the oxygen, a broad program is going to disappoint you for a year, and the better move is to pick the narrowest defensible fight and win it completely. If the field is genuinely split, you have a window that closes as corroboration compounds, and the cost of waiting a quarter is higher than it looks. Either way you now know which game you are in, which is more than most teams running generative engine optimization programs can say. For regulated categories like fintech, where the finance concentration number suggests an unusually open field, that window is the whole opportunity.

You are not behind. You may just be measuring a category average that has nothing to do with your actual competitive position. Find your number first. The strategy falls out of it.

See where you are cited today

A free snapshot audit of your rankings and AI citations before we ever talk.

TT
Tyler TruffiMANAGING PARTNER, SOMETHING INC.

Tyler leads work at the intersection of SEO and generative engines at Something Inc., helping B2B brands get ranked and cited across every major AI engine.

Free consultation

Let us be the last SEO agency you ever work with

A 30 minute call and a free audit of your SEO and GEO position. You keep the findings either way.