How Smashed Lemon's Shopping Agent Users Convert 10.4x Higher
Smashed Lemon is a Dutch women's fashion brand, selling online since 2010 with seasonal collections of tops, knitwear, jackets, pants, dresses and accessories. The brand had been running Shopify's native search and already sorted its collections with the Kimonix Collection Sorting app when it added Kimonix Agentic Search, whose Shopping Agent lets shoppers ask for what they want in their own words.
The results came fast. In the first month, sessions with the Shopping Agent converted at 10.4x the rate of sessions without it (7.3% vs 0.7%).
10.4x
CVR in Agent vs. Non-Agent Sessions
The Challenge: Outgrowing Shopify's Native Search
Smashed Lemon had been using Shopify's native search, which works well while a catalog is simple: it matches the words a shopper types to the product data. A fashion catalog that changes every season needs more than that. Shoppers who aren't sure what to type have no one to ask, and the order of results can only be changed by boosting products by hand, one search term at a time, so keeping each new collection and the sizes in stock at the top meant constant manual work.
The Solution: AI Search Ranked by Merchandising Rules
Kimonix Agentic Search replaced the native search with two stages. Typing in the search bar returns instant AI search results with filters and add to cart. For anything harder to put into keywords, shoppers switch to the Shopping Agent and describe what they want, and it recommends matching products and adds them to the cart inside the conversation. The agent speaks the shopper's language, and synonym lists can map everyday words to the ones the catalog uses.
Because Smashed Lemon runs both Kimonix apps, its search is connected to its merchandising. With the Kimonix team, the brand built a strategy in Collection Sorting, the same app that orders its collection pages, and applied it to the search results and the Shopping Agent's recommendations. Every result now blends the shopper's intent with that strategy: products that sell and convert rank higher, products with core sizes in stock are favored, and anything published in the last 30 days is pushed to the top.
Explore the Kimonix AI shopping agent →
“We've been using Kimonix AI Search for about a month and are very happy with the results so far. The experience has been positive, and communication with Daniel has been excellent. He is responsive, helpful and a pleasure to work with. We look forward to continuing to use Kimonix.”
Laura Muñoz
Smashed Lemon
Smashed Lemon now runs its collections and its search on Kimonix, the all-in-one product discovery platform for merchandising, personalization and search. Each season's strategy is set once in Collection Sorting and shapes both the collection pages and what the Shopping Agent recommends.
Conclusion
“The best Merchandising app there is. Have tried a several ones but Kimonix is by far the best one. And its not just merchandising but have other great features as well. Team is super supportive and customer support is fast.”
Laura Muñoz
Smashed Lemon
Key Takeaways
10.4x
CVR in Agent vs. Non-Agent Sessions
In its first month, Smashed Lemon's Shopping Agent sessions converted at 10.4x the rate of the rest of the site: 7.3% vs 0.7%.
- Shoppers who talk to the Shopping Agent buy at a far higher rate than the rest of the site
- In fashion, search has to follow the season. Applying a Collection Sorting strategy to Agentic Search kept new arrivals and in-stock sizes in front of shoppers
- Measuring agent sessions separately from the rest of the site gave the team a clear read on what the agent is worth
Frequently Asked Questions
What results did Smashed Lemon get with Kimonix?+
In the first month after adding Kimonix Agentic Search, sessions on Smashed Lemon's store that used the Shopping Agent converted at 7.3%, while sessions that did not use it converted at 0.7%. That makes agent-assisted sessions 10.4x more likely to end in an order. The brand also runs Kimonix Collection Sorting, and the same merchandising strategy ranks both its collection pages and its search results.
Is AI search better than Shopify's native search?+
It depends on the store. Shopify's native search is a good fit for a simple catalog, because it matches the words a shopper types to the product data. As a catalog grows and changes every season, AI search adds two things native search doesn't: a shopping agent that shoppers can talk to in their own words, and results ranked by merchandising rules instead of manual boosts. Smashed Lemon, a Dutch women's fashion brand, moved from native search to Kimonix Agentic Search, and in the first month its Shopping Agent sessions converted at 10.4x the rate of the rest of the site.
How do search and merchandising work together on Shopify?+
When both run on one platform, the rules that order your collection pages can also rank your search results. Smashed Lemon applies a Kimonix Collection Sorting strategy to Kimonix Agentic Search, so every result blends what the shopper asked for with what the brand wants to sell: best sellers and high converters rank higher, products with core sizes in stock are favored, and new arrivals are pushed to the top. Setting it once keeps collections and search in step each season, without boosting products by hand one search term at a time.
How do you measure the conversion rate of AI search?+
Split sessions by whether shoppers used the search, then compare their conversion rates over the same period. Smashed Lemon compared sessions that used the Kimonix Shopping Agent with sessions that didn't: 7.3% against 0.7% in the first month, a 10.4x difference. Shoppers who start a conversation are often already close to buying, so read the gap as how well the agent serves high-intent shoppers, and track it over time.