Skip to main content

What this is

A product catalog view that shows which items are Mentioned vs Not mentioned, plus quick access to the responses that mention each product.

Why it matters

  • “Brand visibility” isn’t enough—AI can recommend competitor products instead of yours.
  • Product-level insight tells you what to fix in content and positioning.
  • It’s the fastest way to improve “recommended product” outcomes.

Where to find it

Analytics → Prompt Analytics → Products

How it works

  • Meridian maps product mentions from responses to your catalog.
  • Each product card shows:
    • Mention status (Mentioned / Not mentioned)
    • A shortcut to drill down: See X responses
    • Product metadata (price, tags)
Products grid showing mentioned status and response drilldown Caption: Products surfaces which items show up in AI answers and where.

How to use it

  1. Filter timeframe and platform/attributes if needed.
  2. Scan for Not mentioned products.
  3. Open See X responses for a mentioned product to learn:
    • Which prompts trigger it
    • Which competitors are recommended alongside it
  4. For “Not mentioned” products, identify adjacent products that are mentioned.
  5. Create actions:
    • Add product FAQs/comparisons
    • Improve category pages
    • Earn off-page mentions for the product name

How to interpret results

  • If a high-margin product is “Not mentioned” → it’s missing from model knowledge → add structured pages and comparisons.
  • If only one product is mentioned repeatedly → you have a “default” recommendation problem → broaden product coverage content.
  • If products are mentioned but never ranked #1 → improve proof points and “why choose this” sections.
  • If product mentions only occur on one platform → platform-specific sources → improve those sources across the board.
  • If “See responses” shows the same competitor winning → inspect their cited sources and replicate the structure.

Common questions / troubleshooting

  • “Why don’t I see Products?” → connect a product catalog first.
  • “Why is a product mislabeled?” → confirm SKU/title matching and aliases.
  • “Why is ‘See responses’ low?” → small sample size; expand timeframe.