Setting up shopping prompts

AI Shopping only reveals the opportunities reflected in the prompts you track. The questions you monitor shape the visibility, competitor, and buying-intent insights available for your products. This guide explains how to build a shopping prompt set that delivers meaningful product-level insights.

​

Step 1: Start with categories

Every product in your catalog is automatically tracked. Peec matches products to the prompts you track, helping you understand when they appear in AI recommendations, where they rank, and which competitors show up alongside them. Categories are the backbone of an effective set of shopping prompts. Before answering, ChatGPT often simplifies a shopping prompt into shorter, category-level searches rather than using your exact wording. Our fan-out research shows these retrieval searches tend toward the category, so category-level prompts line up with how the AI actually retrieves. Product-specific prompts work best when layered on top of that category foundation:

• Small catalogs (up to about 100 products): around 3 prompts per product, folded into the category structure.
• Hero products: Track product-specific comparisons and purchase questions, such as “Nike Pegasus vs Adidas Ultraboost”, to get deeper visibility into high-priority products.

​

Step 2: Start with the core prompt set

Before building out your full prompt grid, make sure each category is covered by a small set of high-value prompts:

• One “best” prompt: “What’s the best [product category] for [key use case]?” This is the broad recommendation question that category leaders should be competing to win.
• One “alternatives” prompt: “What are the best alternatives to [your product or brand]?” This captures shoppers who already know your brand and are comparing their options.
• One comparison prompt per key competitor: Focus on your top 2–3 competitors with questions that reflect real buying decisions, such as “I’m training for my first marathon, should I choose [your product] or [competitor product]?”

These prompt types are worth including because they reliably generate either a product shortlist or a side-by-side comparison. The comparison format is especially valuable, since it’s where much of the product attribute data comes from. For a catalog with 5 categories and 3 key competitors, that’s already 25 high-value prompts before you’ve had to make a single strategic choice.

​

Step 3: Think in dimensions

Once your core prompts are in place, expand coverage across three dimensions: These are starting points, not ceilings. A supplement brand with 5 categories, 3 buyer contexts, and 2–3 intents starts around 80–120 prompts; a larger fashion catalog with 15–20 subcategories, around 120–250. Coverage scales from there with every genuinely distinct buying situation you track: more categories, buyer contexts, competitor comparisons, and markets each add new, trackable prompts, and large catalogs comfortably run into the hundreds or beyond. There’s no upper limit on useful coverage as long as each prompt is a distinct buying situation rather than a reword. Categories become Topics; buyer contexts become Tags. So if you have room for more prompts, don’t spend it on rewording. Invest it in broader coverage: more buying situations, more use cases, more competitor comparisons, and more markets. Three simple rules make this framework work:

• Every prompt should represent a unique intent or buying situation: Similar prompts tend to produce similar results, so rewording rarely adds new insights.
• Every prompt should include a clear product category: AI shopping recommendations are typically triggered by specific product searches. (e.g., “best Pilates socks for beginners” is more likely to trigger shopping results than “what do I need for my first Pilates class?”)
• Lean heavily toward commercial intent: Shopping recommendations appear on more than half of commercial prompts, compared to roughly one in ten informational prompts. Keep some informational coverage, but focus your shopping prompt set on product discovery and purchase decisions.

​

Best practices

• Start with the core prompt set: “Best of,” “alternatives,” and “comparison” prompts consistently generate product shortlists and side-by-side comparisons. Those comparison tables are especially valuable because they’re a key source of product attribute data. By including these prompts for every category, you’ll establish a strong foundation of a reliable signal before experimenting with more specialized prompts.
• Keep your prompt set fresh: Some prompts will consistently generate shopping insights, while others won’t. As data comes in, replace low-performing prompts with new buyer contexts instead of letting them sit unchanged.
• Don’t try to solve large-catalog visibility with more prompts alone: The products that appear most often in AI Shopping recommendations are typically the ones already ranking strongly in Google Shopping. That’s why feed optimization is often a more powerful lever than expanding your prompt set.