How would you design an AI shopping assistant to handle vague emotional queries such as 'something my mom would love for her birthday under ₹2000'?

1. Understand query complexity (vagueness). 2. Decompose into structured signals (recipient, occasion, budget, sentiment). 3. Build context based on user history. 4. Map emotional intent to product categories. 5. Rank by giftability, not just relevance. 6. Generate a conversational response. 7. Learn from user interactions to improve the model.