Ulta paid to survey the first children raised on chat assistants, and the useful finding is not the one the trade press ran with. The comfortable read, which most outlets took, was that Gen Alpha still wants stores. The result that matters is quieter: this cohort no longer treats shopping AI as a novelty, and a feature nobody finds novel cannot be sold as a difference.

The study is Ulta’s own, which is worth saying before quoting a figure from it. NielsenIQ fielded it across April and May among 522 American Gen Alpha shoppers, along with 500 parents and guardians. The company selling the AI paid for the questions, which is a reason to read the numbers closely rather than to wave them off. Its baseline result is the one to hold onto: 73% of these young beauty consumers already use some form of personalization tool. That figure covers algorithmic personalization broadly — tailored product recommendations, AI-powered search results, interest-based content suggestions — not AI shopping assistants specifically, where adoption is lower and concentrated in one demographic. The distinction makes the conclusion more precise, not weaker: what is already furniture is the underlying infrastructure, and the assistant is being installed into it. That is not a curve a brand can still take credit for climbing.

What makes the tools infrastructure rather than novelty is how they sit inside the old routine: they add a step; they do not replace one. Nearly eight in ten of these children discover beauty through an online source, and most still want to confirm the find in a store before buying. Those who use AI assistants, in fact, visit stores to browse and try at a higher rate than those who do not, 57% against 36%. A novelty replaces the thing it improves on; a utility gets folded into it, and these shoppers folded it in. The store did not lose to the assistant; it acquired one.

The satisfaction data points the same way, toward usefulness and away from presence. Gen Alpha shoppers who use AI assistants are far likelier than non-users to call personalized beauty relevant to them, 44% against 24%. That gap has two readings, and an honest analyst holds both until one wins: either the tool is working, or the already-keen are selecting themselves into it. The sturdier number concerns trust, where these users are markedly more willing to believe a recommendation once the system explains why it made one, 38% against 23%. Trust tracks the explanation, and an explanation is something a tool either supplies or fails to. Presence was never going to be the variable.

For a brand, utility means a short list of measurable jobs. The assistant either cuts a crowded category to a few defensible picks, or it does not. It either explains the pick in terms a twelve-year-old accepts, or it forfeits the trust the same survey shows explanation earns. An assistant that cannot say why it chose a serum is, to this cohort, a shelf someone else stocked. The work moves from the announcement to the recommender, and the recommender is judged the way a salesperson is judged: by whether the next thing it hands you is the thing you wanted.

Ulta has, in effect, paid NielsenIQ to document the expiry date on its own press releases.

Ulta is an unusually clean case of the reflex the survey quietly undercuts. In April the company shared a stage with Google to launch a Gemini assistant drawing on more than 46 million member profiles, the spring’s most heralded beauty-AI debut, pitched as transformation. By June its own svp of e-commerce described that same system, more soberly, as having learned the store’s opening hours. That announcement addressed today’s shoppers, not the children NielsenIQ was polling, so the charge is not hypocrisy. It is that the reflex behind it — stand on a stage, say the word AI, collect the differentiation — returns nothing from a generation for whom the word describes the weather.


The strongest objection is that these numbers measure selection, not usefulness. If the Gen Alpha children who already find beauty engaging are simply the ones who reach for AI, the satisfaction gaps reveal nothing about what the tools deliver; they only sort the keen from the indifferent. On that reading “win on utility” is the wrong lesson, because no feature is good enough to convert a child who never opens the assistant. The argument fails under one precise condition: if AI use is only a marker of prior enthusiasm and never a cause of it.

Grant the objection in full, and the strategic conclusion survives it. The claim was never that AI manufactures enthusiasm; it was that AI presence has stopped differentiating, and selection makes that truer, not weaker. If the keen self-select into the tools and the rest ignore them, then announcing AI reaches only shoppers for whom it is already table stakes and moves no one else. Either way the announcement buys nothing. What an enthusiast still rewards is the assistant that narrows a thousand serums to three she trusts, and that is a question of utility however she came to open it.

The forecast is narrow and conditional. AI shopping assistants remain a minority habit even among these children, led by teen boys at 26%, nearly double the rate of the other groups. Should that share climb as the cohort ages into its own money, the marginal return on saying “we use AI” falls toward zero, while the return on AI that measurably narrows a choice does not. Beauty has spent two years pricing AI as a headline. The generation forming its habits now will price it as a feature, and features are judged on whether they work. The brands still treating the assistant as an announcement are optimizing for the last audience that will ever be impressed by one.