Émolo (EMOLO FEEL SL) is a Spanish fine jewelry and high-end costume jewelry house founded in Murcia. Every piece is made by hand in its own workshop.
The company operates two channels: direct-to-consumer at emolo.es and wholesale to independent jewelers and retailers across Spain.
Challenge
Search only surfaces what shoppers already know to look for
Émolo faced a fundamental merchandising challenge: how do shoppers discover unique, unexpected items when search only surfaces what they already know to look for?
Our pieces are unique, and if a customer just searches for a generic "gold ring," or deep-links to a single-product landing page, they miss our best work.
Solution
Seller’s Agents that intelligently surface what buyers don’t know to ask for
Product Genius went beyond traditional product discovery tools like Search and Recommendations, bringing the power of AI directly into each shopper journey by pairing an AI seller’s agent with each individual shopper. Just like the clienteling at a luxury jewelry store, these agents learn the customer in real-time.
Seller’s Agents learn from every scroll, hesitation, click, and linger, as well as any other information available. But they don’t wait for a shopper to ask for support. Instead, they use this information to personalize the site for each individual shopper, delivering a TikTok style feed of ideas tailored to that shopper’s in-the-moment behavior.
Product Genius changed how our site feels. It gives shoppers a feed that acts like TikTok—watching every little hesitation and instantly showing them the pieces they didn’t even know they were looking for. It does exactly what our best in-store sales associates do.
Product Genius fits into an existing site user experience, rather than disrupting it. As compelling as AI chat can be at sites like ChatGPT, it can be disruptive to the buyer flow. So it was important to Émolo that Product Genius didn’t require the shopper to click on a chat box to be effective.
Only 1 in 400 shoppers will chat with a chatbot on a shopping site. 100% of shoppers scroll. AI, if it learns quickly enough, can bring in the perfect products and content at the right time, all without being asked. We were thrilled to partner with the Émolo team to prove our new AI in a new market and a new language, using nothing more than the Émolo’s own content.
Results
Adding a new selling surface: The Scroll
This Product Genius Scroll adds a new personalization surface. Placed on a typical PDP, it enables both anonymous and identified shoppers to expand outwards and experience the full breadth of Émolo’s store. The results: in their A/B test, Émolo saw a 15.0% revenue lift on mobile and 27.3% on desktop, creating new revenue when Product Genius interacted with their customers.
Product Genius Seller’s Agents figured out how Émolo's customers actually shop
The numbers only tell a part of the story. Because Seller’s Agents natively understand behavior, Product Genius requires no manual rule configuration. Within days, they automatically identified that Émolo shoppers prefer building jewelry sets and adapted the feed to show matching pieces. They also recognized that ring sizing anxiety was causing purchase hesitation, autonomously intervening by serving Émolo’s 58-second sizing tutorial directly to users stalling on ring pages.
We make every piece by hand in our own workshop in Murcia, and the hardest part of selling online is that a customer can’t feel that. But we don’t have the time or ability to build and maintain rules for every possible interaction, and typical personalization and recommendations don’t understand our products. So we went looking for an AI partner before it was the obvious thing to do.
What surprised me is that the AI arrived at the same two things we say every day in the workshop: our pieces are designed to be worn together, and the first question about a ring is always the size. Without a single rule, the Seller’s Agents just started showing customers our sizing video automatically. And it keeps learning, adapting at a speed and scale we simply couldn’t have imagined before working with Product Genius.
LIM Technology: The AI revolution that transforms what Personalization delivers
Product Genius is based on a unique, new AI technology called a Large Interaction Model (LIM). The LIM learns 30,000 times faster than existing personalization products built as wrappers around technologies from Google, OpenAI or Anthropic. This enables it to deliver real-time learning and response from signals that would overwhelm other AI approaches. And it means that AI can move from the back-office — configuring legacy tools with a chat interface — right to the front of the site, dynamically curating the shopping experiences that drive sales.
- Learns 30,000 times faster than personalization products built as wrappers around Google, OpenAI or Anthropic models
- Real-time learning and response from signals that would overwhelm other AI approaches
- The end of expensive manual rules and tuning complex infrastructure to adjust results
- On Shopify, the app installs in minutes and was fully configured in a few hours
And this doesn’t just deliver customer-facing benefits. For Retailers, it means the end of expensive manual rules and tuning complex infrastructure to adjust results. For Émolo’s Shopify implementation, that meant the application installs in minutes and was fully configured in a few hours. The Seller’s Agent cockpit kept Émolo's team in the loop at every step, ensuring that any decisions were transparent and in line with their priorities, and giving Émolo the power to partner with the AI.
I was bracing for an IT headache. Instead, Product Genius installed in minutes and we had it completely configured in a few hours. My team has loved the leverage Product Genius provides, letting them easily understand its decisions and empowering them to inform the AI of important events like flash sales or holiday spotlights. Product Genius feels like a real teammate for our business.
Revenue lift measured in an A/B test on Émolo's Shopify storefront, with mobile and desktop measured separately.