From Static Catalogs to Conversational Shopping Experiences
An AI shopping assistant is a conversational, software-based retail AI technology that uses generative models and retail data to guide shoppers through discovery, comparison, and purchase decisions in a more humanlike, interactive way than traditional search or menus. Instead of typing rigid queries or clicking through endless filters, customers describe what they want in natural language and receive tailored guidance. These assistants combine search, recommendation, and visual tools such as virtual try-on technology in one interface, aiming to replicate the helpful store associate inside a browser or app. The shift marks a move from simple FAQ chatbots toward agentic AI: systems that can coordinate multiple tools, understand context, and complete multi-step tasks such as refining options, suggesting alternatives, and supporting post‑purchase needs, all while staying inside a brand’s own digital environment.
Kmart’s Joy: An Agentic AI Shopping Assistant for the Home and Wardrobe
Kmart’s new AI shopping assistant, Joy, shows how far conversational shopping experiences have advanced. Built with Google Cloud’s Gemini Enterprise for Customer Experience, Joy is designed as an agentic assistant that spans the full customer journey, from product discovery to post‑purchase support. On Kmart’s site and app, shoppers can use natural language to ask for items by size, style, colour, and budget, then refine results in seconds. Joy also integrates virtual try-on technology so customers can see how selected products may look on them, and a “See It in My Space” feature that places furniture or décor inside photos of their own rooms. Shoppers can upload images to get personalised recommendations and compare options side by side across Kmart, Target, and marketplace brands. As Kmart’s chief customer officer notes, “Customers aren’t just searching anymore; they’re engaging conversationally and looking for ideas and guidance.”
Amazon’s Agentic Shopping Assistant Becomes a Retail Product
Amazon is turning its internal AI shopping technology into a commercial service through AWS, offering retailers an Agentic Shopping Assistant that can be deployed on their own sites and apps. Built on Amazon Bedrock, AgentCore, and OpenSearch, the solution provides starter code, architecture, and expert guidance so retailers can launch AI shopping assistants in weeks rather than years. According to Amazon, more than 300 million customers used its AI shopping assistant last year, and the system was validated on “billions of real shopping interactions on Amazon.com.” That experience is now packaged so brands can build agents tuned to their catalogues, customer segments, and brand voice. Importantly, retailers keep control over proprietary data and relationships while tapping into a technical foundation refined through Amazon’s work on Alexa for Shopping and its newer shopping agent capabilities.
Kate Spade’s AI Gift Concierge and the Rise of Use-Case-Specific Agents
Kate Spade New York illustrates how retailers can adapt shared retail AI technology to a focused use case. Using Amazon’s Agentic Shopping Assistant on AWS and Anthropic’s Haiku 4.5 model, the brand created an AI Gift Concierge that concentrates on one of the most stressful tasks for many shoppers: buying gifts. The assistant holds natural-language conversations, asking about the occasion, recipient, and style preferences before recommending options from Kate Spade’s catalogue. Amazon explains that the design was informed by questions customers have asked Alexa for Shopping and the answers that led to successful outcomes. After roughly two and a half months of testing, Kate Spade launched the conversational shopping experience in April. Early deployments like this suggest AI agents will often start as specialised helpers—gift concierges, outfit builders, or room planners—rather than generic all‑purpose bots.
From Chatbots to Agentic Retail AI Technology
Together, Joy at Kmart and Amazon’s Agentic Shopping Assistant signal a broader shift toward agentic AI in ecommerce. These systems do more than answer FAQs; they handle complex interactions that span product discovery, evaluation, and decision‑making. Shoppers can ask open‑ended questions, upload photos, view items through virtual try-on technology or in‑room visualisations, and compare suggestions in context. The assistants respond conversationally while coordinating multiple tools behind the scenes. For retailers, this promises higher engagement and more informed purchases, as AI shopping assistants reduce friction in finding the right product or gift. According to Amazon, its AI shopping assistant “generated nearly US$12 billion (approx. RM55.2 billion) in incremental sales” in one year, underscoring the commercial stakes. As more brands adopt these platforms, the standard ecommerce interface may evolve from static pages into continuous, two‑way dialogue with autonomous shopping agents.
