Nestlé México's recent La Lechera campaign achieved an open rate nearly twice as high as its previous campaign baseline, a result the brand attributes to AI-driven personalized targeting built on unified customer data. Norma Martín, La Lechera's marketing manager, disclosed the figure during a panel at Treasure AI's Agentic World event in Miami on Tuesday.

The scale behind that result is unusual. Nestlé México claims a presence in 90% of households nationwide and owns more than 70 brands, spanning coffee, candy bars and baby food. That reach is an opportunity to deepen engagement, said Elsa Vences, Nestlé México's data and personalization manager, but it is also "our greatest challenge."

Why did Nestlé México need to unify its data?

For years, the company analyzed data and built targeting on a brand-by-brand basis. It never used findings from one brand to target customers of another, Vences said. She wanted to "break the siloes" the company had historically operated in.

Three years ago, Nestlé México partnered with Treasure AI, a customer data platform that recently rebranded itself as an "agentic experience platform." In practice, that means AI agents are embedded into every step of its data analysis and campaign planning.

Today, Nestlé México connects its data repository — customer names, email addresses and demographic information — to Treasure AI. The platform flags which customers are duplicates, meaning people buying from multiple Nestlé México brands, and builds a fuller picture of their wider interests, Vences said.

The company's individualized data consists mostly of engagement signals within its CRM, such as links a consumer has clicked. Nestlé México also draws on hashed data from Amazon's data clean room, which reveals purchase trends — for example, which age demographic most often buys a product — without exposing personal information.

What did unified data change for La Lechera?

Using engagement data from past campaigns plus demographic information, Nestlé México identified three main audiences for La Lechera, its sweetened condensed milk brand:

• Traditionalists, who use the product in classic recipes
• Experts, bakers curious to try it in new ways
• Explorers, generally younger consumers seeking a quick, easy way to add sweetness to a dish

The brand then targeted each group with recipes matched to their habits: flan for the traditionalists, a pancake topping for the explorers.

Different brands can relate to the "same consumption moments," so Nestlé México often makes decisions about one brand based on data from several others, Martín said. La Lechera, Coffee-Mate and KitKat all appear in moments of "indulgence," so the brand was able "to feed more users into the La Lechera base" by understanding shared commonalities among users of all three.

How do AI agents help smaller brands?

The harder problem, Vences said, is targeting consumers of Nestlé México's smaller brands, because the company holds less data on those customers and understands less about what they want. At least, that used to be the case.

Nestlé México now determines whom to reach and how by prompting Treasure AI's agents with specific campaign parameters. The agent generates suggestions based on data from relevant brands. Vences said she could ask the agent to build an audience of Gen Z chocolate lovers, and it would pull from data across larger brands like KitKat or from people who engaged with dessert recipes in other brands' campaigns. Brands can also upload a website URL and ask the agent to find recipes on the site that would resonate best with a given audience.

For years, creative agencies have had to do this work manually. With AI, Vences said, it is not manual, and it is not just based on intuition. Nestlé México's targeting is now "more dedicated," she said, using stronger data to reach customers more accurately and holistically.