An SMS sales assistant for a wine retailer

The problem

A wine retailer was struggling to convert customer inquiries into sales. Their catalogue had hundreds of SKUs across different varietals, regions, and price points, too complex for customers to browse manually, and too time-consuming for the sales team to handle one-on-one at scale. They needed a way to let customers self-serve discovery and ordering without building an app nobody would download.

What we built

We built a conversational AI assistant that worked entirely over SMS: no app, no login, no friction. Customers could text in natural language, such as a grape variety, a flavor profile, a price range, or just “something under ₹2000 for a dinner party”. The system would parse their intent, search across the full SKU catalogue, and return the top 3 most relevant matches with details.

Customers could then place an order directly by replying to the message. The order was captured and routed automatically to the sales team's dashboard for fulfilment, with no manual data entry and no missed inquiries.

The result

The sales team went from manually fielding individual product questions to receiving structured, ready-to-act orders. Customer discovery went from a friction-heavy browsing experience to a 3-message conversation. The channel required zero onboarding for end users: if they could text, they could use it.

Tech approach

Natural language processing for intent and entity extraction, integrated with product catalogue search, an SMS gateway, and the sales team's order management workflow.