The serum bottle sits on a surface the color of wet sand, lit from the left, a single bead held on the dropper. You have seen it before, or seen something calibrated to within a Pantone chip of it. L’Oréal’s content engine doesn’t make that picture once; it manufactures the supply of it, thousands of near-identical frames built to brand guideline and pushed wherever a feed will take them. The group’s expanded deal with OpenAI, announced June 17 at VivaTech, bets that beauty marketing now means flooding the surfaces most searches now leave behind, until the brand itself is the answer the machine reads back.

The headlines went to the showroom. Maybelline’s virtual try-on moves inside ChatGPT on L’Oréal’s ModiFace engine, so the shade lands on your own face mid-conversation. Lancôme and Kérastase get tuned product discovery in the US; SkinCeuticals, CeraVe and Garnier already run in OpenAI’s global ChatGPT ad pilot. The engine behind all of it is CreAItech, L’Oréal’s in-house generative platform, now running on OpenAI’s newest model. The company built it, Digiday reports, for “earned, owned and paid social, e-commerce and other environments that require vast quantities of images.”

CreAItech has already cut production costs by 40% and turned out 50,000 marketing assets, work the group’s 10,000 marketers once commissioned a shot at a time. Another 73,000 employees have been trained on generative tools. Set that against an ad budget that rose 10% last year to $15.48 billion, and the rationale is plain: when distribution is infinite and nearly free, the creative has to be too.

The surfaces it feeds have stopped sending traffic back. By the first four months of 2026, 68.01% of US Google searches ended without a click, up from 60.45% two years earlier — SparkToro notes the two figures draw on different clickstream panels and aren’t directly comparable, so read the direction, not the gap. AI Overviews appear on more than a fifth of searches and cut clicks by about 60% where they show, an Ahrefs figure carried in the study and reported by Search Engine Land. Rand Fishkin’s read was that brands should win attention where audiences already are, click or no click; infinite creative is how a company L’Oréal’s size does it.

This is the no-click sell, and L’Oréal wrote its first draft in 1971.

That year, a 28-year-old copywriter named Ilon Specht wrote “Because I’m worth it” for L’Oréal’s Préférence hair color. The line asked for no proof and invited no comparison; the purchase was self-evident, settled before the shopper could look elsewhere. That is the posture of the zero-click answer, a recommendation accepted without a second tab. L’Oréal scripted that sell more than fifty years ago, and has now built the plant to print it by the thousand.

The ModiFace integration has an aesthetic justification; the content factory is the harder case. The shade landing on your own face inside the conversation changes what beauty looks like at the moment of choosing, which is technology earning its place by changing the image. When 50,000 assets all answer to one guideline, the single seductive frame beauty advertising was built on gives way to the unit of inventory, and the reader you design for stops being a human eye and becomes a model that averages.

The bet is sound on its own terms, and the cost won’t show on the balance sheet. Run enough brands through the same engine into the same answers, and the surfaces converge on a house style that belongs to no house. L’Oréal can win the position of the answer and still lose what made a shopper want the bottle: the sense that this one was not like the others.