From keyword search to intent-aware AI shopping assistants
An AI shopping assistant in e-commerce is a large language model–powered system that uses conversation, context, and multimodal inputs to understand shopper intent, recommend products, and personalise the product search experience beyond simple keyword matching or static filters. Instead of typing short queries and scrolling through long lists of results, customers can describe what they need in natural language, upload photos, and refine options through dialogue. Large language models (LLMs) make sense of loose descriptions, such as “something for a small apartment in neutral tones,” and connect them to relevant inventory. This shift is reshaping LLM e-commerce design: discovery starts with a guided conversation rather than a blank search box, and the assistant continuously adapts to preferences, budget, and style as the session unfolds.
How Google and Amazon are redefining product search experience
Major platforms are rethinking how shoppers find products, using LLMs to interpret intent instead of relying on exact keyword matches. Google is weaving generative AI into shopping results so users can ask broad questions, compare options, and see summarised suggestions without crafting perfect queries. Amazon is also experimenting with LLM-driven features that act like a knowledgeable store associate, capable of answering open-ended questions and surfacing personalised product discovery paths based on context from previous interactions. Together, these moves mark a shift from static lists of links to dynamic, conversational guidance that feels closer to asking a human for help. For shoppers, this can mean fewer irrelevant results and faster decisions. For brands, it means the battle for attention is increasingly mediated by AI systems that decide which products appear first.
Inside Kmart’s Joy: virtual try-on and “See It in My Space”
Kmart’s Joy AI shopping assistant shows how LLM e-commerce is becoming more visual and interactive. Joy lets customers try on selected products virtually and use “See It in My Space” to place furniture or décor inside their own rooms before buying, a clear step forward in virtual try-on technology. The assistant supports natural language prompts like “show me a sofa in a dark colour under this budget,” then refines suggestions within seconds. Shoppers can upload photos to receive tailored recommendations and compare items side by side across Kmart, Target, and marketplace brands. According to Bernard Wilson, Chief Customer Officer at Kmart Group, “Customers aren’t just searching anymore; they’re engaging conversationally and looking for ideas and guidance,” and Joy is designed to narrow choices while staying focused on affordable, everyday products.
Why AI-powered discovery is becoming a key retail battleground
As more retailers adopt AI shopping assistants, product search experience becomes a competitive differentiator rather than a basic utility. Whoever offers the fastest, most accurate, and most helpful personalised product discovery gains an edge in shopper loyalty and conversion. Tools like Joy, built with Gemini Enterprise for Customer Experience, highlight how AI can support the entire journey from discovery to post-purchase service with a single conversational interface. Retailers are racing to integrate multimodal inputs, conversational context, and visual previews so shoppers can search by text, image, or both. The result is that what customers see first online is no longer a neutral ranking of products; it is a curated, intent-aware selection shaped by LLMs. In this new landscape, the real contest is to become the default assistant shoppers turn to when they start browsing.
