TL;DR
Built a conversational AI chatbot inside the 7-Eleven Thailand app, powering both customer support and chat commerce for 16M+ monthly active users across 15,000+ stores nationwide. Users could browse products, check real-time stock at nearby stores, get personalized recommendations, and complete purchases, all without leaving the chat.
The Problem
7-Eleven Thailand isn't like 7-Eleven anywhere else. With over 15,000 branches blanketing the country, it's less a convenience store chain and more a piece of national infrastructure. The mobile app had 16M+ monthly active users, and the company wanted to unlock new digital sales channels as part of a larger digital transformation.
The goal was to build a conversational layer inside the app that could serve across three distribution channels: answering customer inquiries, driving product sales through chat, and supporting the app's real-time delivery service. All of it had to work across 15,000+ stores with live inventory, local promotions, and a nationwide user base.
Fun fact
7-Eleven in Thailand is genuinely amazing. We have over 15,000 branches. Seriously, look it up!
Context
This was before the ChatGPT era. Every intent was mapped by hand, every fallback designed deliberately, every response tree built from scratch. No LLM safety net, just NLP, pattern matching, and a lot of conversational design.
The Solution
A chat commerce experience inside the 7-Eleven app, where users could discover, decide, and check out without ever leaving the conversation. The same chatbot also handled customer support (store locator, hours, promos, delivery order status), but the ambitious part was turning a chat thread into a full shopping flow.
Real-time stock at your nearest store
The chatbot detected the user's location, mapped them to their closest 7-Eleven stores, and queried live inventory across 15,000+ stores to show only what was actually on the shelf. No dead ends, no out-of-stock disappointments.
Personalized product recommendations
Product carousels surfaced inside the chat thread, tailored to each user's purchase history and profile, with current promotions applied automatically. Every conversation felt relevant rather than generic.
Complete purchase in chat
The entire shopping flow, from discovery to checkout, happened inside the conversation. No app switching, no separate cart screens, just one thread from the first question to the confirmation message.
The Impact
16M+
monthly active users
15k+
stores nationwide
3
distribution channels
+15%
customer satisfaction