- Lazy GPT wrapper. Same answer for every store.
- No memory of your products, policies, or rules.
- Hallucinates sizes, prices, return policies.
- Speaks like a help desk. Sells nothing.
- Endless config, complicated flows and rules.
- "I'm sorry. I can't help with that." 🙃
Hi, I'm Jugal.
01 — Why I built this
I fired half my support team. The other half got their evenings back.
I run Bombay Trooper (estd. 2013) and Sytara.in (estd. 2019),two D2C brands with 400,000+ customers. My office today runs on AI agents and a tiny ops team.
The problem
- Pre-sale chat drowned us: "Will this fit?", "COD?", "Discount?" on loop.
- Existing AI plugins = GPT wrappers. ~40% hit rate, generic voice.
- Midnight questions = lost sales.
What I did
- Dumped 10+ years of real support tickets into a custom-trained model.
- Wired it up to live Shopify data: prices, stock, policies.
- Added RAG, vector search, and other AI tricks to make it know your products.
- Experimented it on my own stores first until the answers felt like me and not ChatGPT.
Result Half the support team. Higher conversion. Happy customers. CartChat is designed to deliver.
Receipts