Design AI for ecommerce search and product discovery.

### Signal to interviewer

I can design AI commerce discovery that improves both conversion and shopper confidence.

### Clarify

I would clarify catalog structure, shopper intent patterns, and business constraints around merchandising and sponsorship.

### Approach

Build an intent-to-catalog engine: semantic query parsing, relevance + exploration ranking, and decision-assist content on product detail pages.

### Metrics & instrumentation

Primary metric: qualified conversion from AI-assisted journeys. Secondary metrics: add-to-cart quality, browse depth efficiency, and assisted comparison usage. Guardrails: return rate increase, sponsored-result confusion, and recommendation complaint volume.

### Tradeoffs

Higher personalization improves conversion but risks filter bubbles. Merchandising pressure can increase revenue but may reduce perceived fairness.

### Risks & mitigations

Risk: inaccurate product attribute inference; mitigate with source validation. Risk: overfitting to historical clicks; mitigate with exploration controls. Risk: trust loss from opaque ranking; mitigate with recommendation explanations.

### Example

For home appliances, AI narrows options by kitchen size, energy preference, and delivery window, then surfaces tradeoff comparisons across shortlisted products.

### 90-second version

Design ecommerce AI around intent clarity and confident decisions. Optimize qualified conversion while guarding transparency, post-purchase satisfaction, and fair ranking behavior.