Your shopping cart is empty!
WordLift merged four L'Oréal brands into one knowledge graph for AI search
Publication date: 29.09.2026
In short
- WordLift built a shared knowledge graph architecture for four L'Oréal brands in Türkiye — L'Oréal Paris, Garnier, Maybelline New York and NYX Professional Makeup.
- Product pages were turned into the source AI systems cite, instead of optimizing each brand separately.
- WordLift's own blog claims up to 450% growth in AI search visibility, but that's a vendor's own number, and there's no independent audit behind it.
WordLift, an Italian company that has spent years building knowledge graphs and schema.org markup for large catalogs, explained on its blog how it brought four L'Oréal brands in Türkiye under one shared data architecture.
Four brands, one knowledge graph
The playbook is simple: instead of building a separate SEO strategy for each brand from scratch, WordLift tied products, ingredients, shades and categories into one entity structure. The same approach was applied to L'Oréal Paris, Garnier, Maybelline New York and NYX Professional Makeup.
When the data structure is consistent, an AI system can more easily recognize that a Maybelline mascara page and a Garnier shampoo page are the same cosmetic-product entity type with different attributes. According to the company, pages built this way started showing up as sources in AI Overviews and chatbot answers.
Technically, it works like this: product, brand and offer pages get tied together through schema.org — Product, Brand, Offer — with shared attributes across all four brands, like skin type or product purpose. An AI system reads the whole node of connected entities around a product and builds its answer from that, without being tied to a single page.
up to 450%WordLift's own estimate of how much AI search presence grew for one of the four brands after the knowledge graph rebuild
A number from a vendor's own blog
Worth pausing here. WordLift sells knowledge graph services, and the L'Oréal Türkiye case study runs on its own blog — it isn't an independent study or a third-party assessment. WordLift doesn't disclose the methodology: it's unclear which queries, which AI systems, or what time period went into the sample. For example, “up to 450% growth” could mean going from two mentions to eleven in a month, or a citation share moving from 1% to 5.5% — both are honestly “up to 450%,” yet they describe completely different scales.
We've covered how schema.org markup and structured data help a page become a source for AI in general, and markup alone doesn't guarantee anything without solid content behind it.
For a market where one retailer often runs several sub-brands at once, the format itself is the interesting part: not four separate SEO teams, but one architecture scaled across brands. WordLift laid out the full case on its own blog.
What struck me was less the 450% figure and more the fact that a major brand is willing to show its knowledge graph architecture in public — details like that usually stay behind closed doors. I suspect WordLift's competitors will publish similar case studies soon, with numbers just as hard to verify from the outside.
