Challenge
Gold jewellery shoppers needed instant answers on live rates, product discovery, comparisons, and return policies — without waiting for support or navigating complex catalog filters.
ai · 2025
AI chatbot for NKM B2C gold jewellery — live gold rates, smart product suggestions, side-by-side compare, and instant policy answers.

Overview
The NKM AI Shopping Assistant is a production eCommerce chatbot built into the NKM B2C web and mobile apps — powered by AWS Bedrock Converse with tool calling, a jewellery knowledge base, and bilingual quick replies in English and Hindi.
Challenge
Gold jewellery shoppers needed instant answers on live rates, product discovery, comparisons, and return policies — without waiting for support or navigating complex catalog filters.
Strategy
Orchestrate Bedrock with structured tools (search, recommend, compare, gold rate, FAQ) and a knowledge base for policies and care guides — surfaced through quick-reply chips on web and Flutter mobile chat.
Solution
Laravel ChatOrchestrator with BedrockConverseService, ChatToolExecutor for catalogue actions, and ChatQuickReplies for one-tap intents — delivering product cards, live 22K/24K rates, compare tables, and detailed policy answers in chat.
UX note
Priority was given to readable hierarchies, status visibility, and journeys that support busy users — not decorative complexity.
Technology
Process
Features
Results
Architecture
Visuals
Selected visuals from the delivery set, used to communicate product direction and interaction density.

Screen 01
NKM AI Shopping Assistant

Screen 02
NKM AI Shopping Assistant

Screen 03
NKM AI Shopping Assistant
Testimonial
Feedback from the team we built this project with.
“Customers ask about gold rates, product options, and return policy all the time. The AI chatbot answers those questions instantly on web and mobile — it saves our team hours and helps shoppers decide faster.”
Related projects
Next step
Share your industry context and goals — we'll help frame scope, UX priorities, and architecture.