THE BRANDS EVERYONE KNOWS, CHOSEN BY NOBODY

Eleven oral care brands. Four AI models. Fifteen questions a real buyer would actually ask, from the best everyday toothpaste to what to do about chronic bad breath. Sixty moments where a machine hands somebody an answer.

The category got better this quarter. The median score rose five points. Almost every brand that was behind moved up. And the four brands that led the category in April all went backwards relative to it.

This is the second ARA reading of oral care, which makes it the first with movement in it. One quarter is a photograph. Two is a direction.

The thesis, stated once, plainly: every household name in this category is understood by the machines. Almost none of them is chosen by one. Colgate, Crest and Sensodyne all score the maximum on AI recommendation. Between them they take the first answer in twenty of sixty moments. Crest takes three.

ABOUT THE RANKING

Five attributes, each scored 0–20, summed to 100.

A1 Findability — can machines find and parse you.

A2 Clarity — do they understand what you are.

A3 Sentiment — what the culture remembers.

A4 Identity Fidelity — is your personality distinct enough to reproduce.

A5 AI Recommendation — when a buyer asks, are you named.

Trust Tier: AWESOME (83–100) / STRONG (70–82) / AVERAGE (56–69) / WEAK (40–55) / INVISIBLE (0–39). The five attributes sum to a Trust Tier — a single read on how far the machines will go on your brand's behalf.

Position and movement are measured differently. Position is scored on the full current question battery. Movement is scored like-for-like, only on the questions carried forward word for word from Q2 — the only basis on which two quarters are the same measurement. Where the two disagree for a brand, both numbers are correct and they are answering different questions.

The instrument note that matters: the question set was widened this quarter from eight questions to fifteen. Movement is reported only on the eight carried over word for word. The fifteen-question figures are this quarter's standing, not this quarter's movement. All four models were reading the live web for this reading, which was not true of every earlier one.

THE ORAL CARE LEADERBOARD

CATEGORY CONTEXT / ORAL CARE Q3 2026 / 11 brands / median 66 / spread 80→46 (34 pts) / second reading, every brand carries a movement figure

Chart 01 — The Q3 oral care leaderboard. Eleven brands, each with its Trust Tier named beside its score and its movement against Q2.

A 34-point spread, and a reshuffle at the top. Oral-B takes the category at 80. Sensodyne, which led in April, is second at 79 having lost a point. Crest is third at 73, down two. Five of the six brands in the bottom half gained six or seven points.

Only one brand moved badly, and it moved a long way.

KEY FINDING 01

THE CATEGORY ROSE AND ITS LEADERS FELL

The median brand gained five points. Measured against that median, the standings invert:

Up: Arm & Hammer 54 to 61. Parodontax 61 to 68. Marvis 51 to 57. Tom's of Maine 60 to 66. Hello 56 to 62. Down, once the category's own rise is taken out: Colgate, Listerine, Sensodyne, Crest.

Sensodyne lost a point and Crest lost two, which sounds like noise. Against a category gaining five, it is six and seven points of ground. Standing still in a rising category is a decision to be overtaken slowly.

KEY FINDING 02

BEING NAMED IS NOT BEING CHOSEN

Colgate is named in twelve of the fifteen questions. Crest in eleven. On paper they are the two most present brands in AI oral care.

Colgate converts that presence ten times in sixty answers. Crest converts it three. Both score the maximum on AI recommendation, because that measure rewards being on the list at a good position, and both are reliably second. Second is not a smaller version of first. It is a different outcome: the buyer reads one name and stops.

Where Colgate does win it wins outright. Whitening, first across all four models. Everyday toothpaste, first in three of four. Sensodyne does the same on a narrower base: sensitive teeth in every model, enamel in three. The pattern is that first place goes to brands with one claim, stated the same way for a very long time.

Chart 02 — Known vs chosen. Brand fundamentals against recommendation presence. Crest, Listerine and Arm & Hammer are indistinguishable on fundamentals and are named at 20, 16 and 8.

KEY FINDING 03

THE SHELF GOT LONGER AND THEY DID NOT GROW WITH IT

One hundred and thirty-five brands were chosen at least once across the fifteen questions. The eleven audited here, which between them hold most of the physical shelf in most of the world, hold 41.6% of this one.

Split that by question. On the eight carried word for word from April, they hold 48.2%. On the seven newer ones a buyer actually asks now — kids, bad breath, what is worth paying more for, what belongs in your bathroom — 33.6%. The older the question, the better the old names do.

Measured on those eight identical questions, every one of the eleven lost share this quarter. Crest gave up the most, then Colgate, then Tom's of Maine. Not one gained. Between them they went from holding 63.1% of those answers to 48.2%.

One caveat is worth stating plainly, because it cuts against our own headline: this is the first oral care reading in which all four models searched the live web, and answers ran deeper. Better-informed models find more real brands, so some of that challenger gain is the instrument improving rather than the incumbents slipping. The direction is not in doubt. The precise size of it is.

Chart 03 — The Shelf. Every moment an AI hands a shopper a recommendation: 15 buyer questions × 4 models = 60 moments. Tile size is position in the answer. Toggle between the audited cohort and the full shelf.

The full shelf — all 135 brands ranked and annotated, and the challengers now sitting mid-shelf among the majors — is in ARA AI Shelf: Oral Care Q3 2026.

KEY FINDING 04

NOTHING HERE IS HIDDEN

Across the eleven brands, not one is a brand whose strength the machines are failing to see. Six are read accurately for what they are. Five are read as weaker than they were a quarter ago, and are.

That is an unusual result and an expensive one. In most categories we read, at least one brand is carrying real equity the machines have not caught up with, and the fix is translation. Here there is nothing to translate. Where these brands are missing from an answer, they are missing because there is no evidence in the world for a machine to find.

Chart 04 — Understood vs distinct. Each line is one brand: Clarity against Identity Fidelity. Clarity runs at a median of 17; identity at 8.

KEY FINDING 05

THE CATEGORY LEADER IS A HARDWARE BRAND

Oral-B tops this reading at 80, and appears in six of the fifteen questions. It wins the electric toothbrush answer, the what-is-worth-paying-more-for answer, and a share of the dentist answer. In the toothpaste aisle, where most of the category's revenue is, it is close to absent.

A narrow, decisively owned position now outranks a broad, softly held one. That is the single most transferable finding in this reading.

KEY FINDING 06

ONE BRAND FELL OFF THE SHELF

Quip went from 62 to 46, twenty-one points below the category once its rise is accounted for. It is now named in five of sixty answers and takes the first answer in none.

Quip was built as a subscription brand with a design story, and both are hard for a machine to verify. When the recommendation surface asks what actually works, a brand whose strongest asset is how it looks and how it ships has nothing to hand over.

WHAT THIS MEANS IF YOU ARE AN ORAL CARE BRAND

The oral care category is not losing its position with AI. It is holding a position it no longer converts. The names are known, the claims are understood, the sentiment is fine. The answer still goes somewhere else about a third of the time, and on the newest questions two thirds of the time.

The work is not to become better known, and it is not to become better described. Both are already strong across this category. The work is to decide which question you intend to own, and to make your brand the obvious answer to it. Sensodyne has done that with sensitivity for decades and still wins it across every model. Oral-B has done it with hardware and now leads the category from six questions out of fifteen. Every brand below them is competing for the general answer, and the general answer is where the machines have the most names to choose from.

Second place in an AI answer is not a podium. It is a list nobody reads to the end.

ARA Oral Care Q3 2026 / 11 brands / Claude, GPT-4o, Gemini, Perplexity / second reading of the category, question set widened from 8 to 15 and movement reported only on the 8 carried word for word. Top score Oral-B 80; median 66. Your brand isn't listed? Request an ARA audit → index.araco.ai