Rule-based / Mock AI
The ShopSphere assistant is a rule-based assistant. It does not use a large language model or send your messages to any AI service. It understands shopping requests using keyword, pattern and entity matching, turns them into structured filters, and runs those filters through the same search and recommendation services the rest of the store uses.
| You say | It understands |
|---|---|
| “Show me smartphones under 30,000 rupees with good ratings” | { category: smartphones, maxPrice: 30000, minRating: 4 } |
| “Headphones under ₹5k” | { category: headphones, maxPrice: 5000 } |
| “Show me something cheaper” | keeps category, lowers maxPrice to the median of what was shown |
| “Something with better ratings” | keeps category + price, raises minRating |
| “Only in blue” | adds colour: Blue to the current filters |
| “More like the second one” | runs the Similar Products strategy seeded with result #2 |
| “Never mind, show me laptops instead” | resets category, keeps general preferences where they still fit |
If a message can’t be understood confidently, the assistant asks a clarifying question instead of guessing. Conversation history is kept only for your current browser session.