Skip to main content
AI shopping assistants don’t always respond the same way to every query. Depending on the buyer’s intent and the platform, they shift between two distinct response modes: Carousel and Research. Understanding shopping modes helps you interpret your prompt-level data and know what kind of AI content is ranking your competitors. In carousel mode, the AI returns a visual grid of product cards — typically 3–6 products with names, images, prices, and a brief description. This is the AI equivalent of a search results page. When it happens: Purchase-intent queries, where the buyer is ready to buy and wants options fast. Example query: “best trail running shoes under $150” What a win looks like: Position matters enormously here. Position 1 gets the most attention. Being absent from a carousel prompt is a meaningful miss. How to optimize: Strong product listings, structured data, accurate pricing, and review counts are the primary ranking signals in carousel responses.

Research mode

In research mode, the AI returns a long-form buyer guide — a detailed text response that compares options, explains trade-offs, and makes a recommendation. This is sometimes called “deep research” mode. When it happens: Comparison, feature-research, and general-research queries, where the buyer is still evaluating. Example query: “what should I look for in trail running shoes for technical terrain?” What a win looks like: Presence is more important than position here. Being cited in a trusted research response builds brand authority and influences buyers earlier in the funnel. How to optimize: Authority signals matter most — third-party reviews, editorial coverage, and expert citations. Brands that are well-covered by authoritative sources rank well in research responses.

Summary

ModeBuyer stagePrimary ranking signalWhat a win looks like
CarouselReady to buyProduct data qualityPosition 1–3 in a purchase query
ResearchStill exploringBrand authority & coverageNamed as a top recommendation in a buyer guide
Your prompt table shows the mode for each query. Filter by mode to understand where your visibility gaps are most costly.