Gen Z Wants to Be Part of the Journey, Not Just the Destination
CANNES, France — You know the old playbook of showing consumers product ads on white backgrounds and hoping for the best? Apparently, younger shoppers want immersion, personalization, and the thrill of discovery – not a sterile transaction.
With 80 billion monthly searches on its platform – half of which carry commercial intent – Pinterest finds itself at the center of that trend, where inspiration and purchase increasingly blur into a single experience.
“Gen Z, they want to be part of the discovery, meaning that they want it to be immersive, personalized, visual,” said Matt Madrigal, chief technology officer, Pinterest, in this video interview with Beet.TV at the Cannes Lions Festival of Creativity. “The marketers that will do well will recognize that so they can tell stories to Gen Z.”
The limitations of text-based search
Search has changed. While traditional search engines excel when users know exactly what they want, they struggle with the messy, exploratory nature of modern shopping journeys.
“You’re starting to see some of the limitations of text-based search,” Madrigal said. “Text search now is becoming much more conversational, you have longer queries. But with that, there’s no inspiration, there’s no discovery.”
Pinterest positions itself as the answer to that gap. According to Madrigal, more than 40% of Gen Z users come to Pinterest to search first, while 60% use the platform specifically for shopping. “It starts often times with Gen Z wanting to search with an image or an idea,” he said.
Building a taste graph
Central to Pinterest’s approach is what the company calls its “taste graph” – a massive data structure built from user interactions across 750 billion pins organized into 15 billion boards.
“The taste graph is all this incredible user interaction data that we have,” Madrigal explained. “All the associated visual searches, text searches allows us to really understand what a user’s taste or vibe or aesthetic is.”
The company recently unveiled several AI-powered tools to deepen this capability. At Cannes 2026, Pinterest announced its Business Assistant, an AI collaborator currently in closed beta, along with a new Pinterest Model Context Protocol that optimizes ad performance at the asset level rather than the ad level.
Ads as content
The traditional model of advertising as interruption is giving way to something more integrated. Madrigal argued that when ads are genuinely relevant, they function as content rather than disruption.
“We always view really relevant ads as great content,” he said. “When you have great content, it really resonates with our users. And so that delivers both a great engaging experience for our users and usually better outcomes for our advertisers.”
Pinterest’s proprietary recommendation models play a key role in this approach. “They’re 30% more accurate compared to recommendations you’ll see through other commercial models,” Madrigal claimed, explaining why the company builds its own foundational AI models rather than relying on third-party solutions.
Anticipating what users want next
The real ambition goes beyond simple personalization – showing users more of what they have already clicked on. Madrigal described a more sophisticated goal: anticipating desires before users consciously recognize them.
“Traditional personalization has been something where I click on this, therefore I like that,” he said. “What we’re trying to do is, as we try to anticipate what you want next, we do that in a way for users when they don’t even know they want it themselves.”
This capability relies on Pinterest’s logged-in user base of 631 million monthly active users, providing rich behavioral data that powers the platform’s recommendations. The company also recently launched an experimental app called Ask Pinterest, designed to explore conversational, visual-first shopping experiences.
CANNES, France — You know the old playbook of showing consumers product ads on white backgrounds and hoping for the best? Apparently, younger shoppers want immersion, personalization, and the thrill of discovery – not a sterile transaction.
With 80 billion monthly searches on its platform – half of which carry commercial intent – Pinterest finds itself at the center of that trend, where inspiration and purchase increasingly blur into a single experience.
“Gen Z, they want to be part of the discovery, meaning that they want it to be immersive, personalized, visual,” said Matt Madrigal, chief technology officer, Pinterest, in this video interview with Beet.TV at the Cannes Lions Festival of Creativity. “The marketers that will do well will recognize that so they can tell stories to Gen Z.”
The limitations of text-based search
Search has changed. While traditional search engines excel when users know exactly what they want, they struggle with the messy, exploratory nature of modern shopping journeys.
“You’re starting to see some of the limitations of text-based search,” Madrigal said. “Text search now is becoming much more conversational, you have longer queries. But with that, there’s no inspiration, there’s no discovery.”
Pinterest positions itself as the answer to that gap. According to Madrigal, more than 40% of Gen Z users come to Pinterest to search first, while 60% use the platform specifically for shopping. “It starts often times with Gen Z wanting to search with an image or an idea,” he said.
Building a taste graph
Central to Pinterest’s approach is what the company calls its “taste graph” – a massive data structure built from user interactions across 750 billion pins organized into 15 billion boards.
“The taste graph is all this incredible user interaction data that we have,” Madrigal explained. “All the associated visual searches, text searches allows us to really understand what a user’s taste or vibe or aesthetic is.”
The company recently unveiled several AI-powered tools to deepen this capability. At Cannes 2026, Pinterest announced its Business Assistant, an AI collaborator currently in closed beta, along with a new Pinterest Model Context Protocol that optimizes ad performance at the asset level rather than the ad level.
Ads as content
The traditional model of advertising as interruption is giving way to something more integrated. Madrigal argued that when ads are genuinely relevant, they function as content rather than disruption.
“We always view really relevant ads as great content,” he said. “When you have great content, it really resonates with our users. And so that delivers both a great engaging experience for our users and usually better outcomes for our advertisers.”
Pinterest’s proprietary recommendation models play a key role in this approach. “They’re 30% more accurate compared to recommendations you’ll see through other commercial models,” Madrigal claimed, explaining why the company builds its own foundational AI models rather than relying on third-party solutions.
Anticipating what users want next
The real ambition goes beyond simple personalization – showing users more of what they have already clicked on. Madrigal described a more sophisticated goal: anticipating desires before users consciously recognize them.
“Traditional personalization has been something where I click on this, therefore I like that,” he said. “What we’re trying to do is, as we try to anticipate what you want next, we do that in a way for users when they don’t even know they want it themselves.”
This capability relies on Pinterest’s logged-in user base of 631 million monthly active users, providing rich behavioral data that powers the platform’s recommendations. The company also recently launched an experimental app called Ask Pinterest, designed to explore conversational, visual-first shopping experiences.
Series Videos
