Ecommerce Voice AI — Product Search, Comparison & Cart

A voice shopping assistant that connects a customer's request to product search, comparison, and a confirmed demo cart.

THE PROBLEM Choosing a desk setup means balancing budget, specifications, and personal preferences across several products. I built this interactive ecommerce demonstration so a shopper can explain what they need, ask follow-up questions, and see the recommendations update as they speak.

WHAT I BUILT The LiveKit voice agent searches a sample catalog before recommending products. Tool calls update the Next.js interface with matching items, side-by-side comparisons, and a bundle total calculated from catalog prices. When the shopper changes their requirements, the agent refreshes the search rather than relying on stale recommendations.

CONTROLLED CART ACTIONS The agent prepares the selected bundle, reads the total, and requests confirmation. A separate customer turn is required before the demo cart changes. The application checks that the confirmed items match the pending bundle.

MY ROLE I built the shopping interface, voice-agent workflow, catalog tools, bundle calculations, and cart-confirmation logic using Next.js, TypeScript, LiveKit, and LLM tool calling.

TRY THE DEMO https://ashirshaikh.com/voice-shopping-demo

This is an independent demonstration using sample products and a demo cart. It does not process payments or place real orders. A live store deployment would require its own catalog, commerce integration, and operational testing.