Search Intent Agent
Maps consumer search language against a merchant catalog. Surfaces the exact keywords and natural-language questions consumers use, flags ghost products (catalog entries invisible to high-demand searches), and recommends title and landing-page rewrites that match consumer intent.
Endpoint: POST /a2a/intent-agent
Skills
Ghost Product Detection
Identifies products in the merchant catalog whose titles do not match the high-demand keywords consumers are searching for.
Example queries:
- "Which of my products are invisible to search?"
- "Find ghost products in my kitchenware catalog"
Keyword Landscape Mapping
Maps the full keyword landscape around a category by aggregating seed, canonical, and intent keyword types.
Example queries:
- "Map the keyword landscape for home fitness"
- "What are the top search terms in premium yoga?"
Voice & Natural-Language Question Mining
Surfaces natural-language questions consumers ask in voice and AI search interfaces.
Example queries:
- "What voice-search questions are trending for kitchen knives?"
- "Show me natural-language questions about yoga mats"
Product Title Optimization
Suggests product title changes grounded in consumer language and high-SV keyword data.
Example queries:
- "Rewrite my product titles to match consumer language"
- "Which titles should I update for better visibility?"
Modality Analysis (Text vs Voice/AI)
Breaks down opportunity by text search vs voice/AI search modality so merchants can invest accordingly.
Example queries:
- "How much of my demand is voice vs text?"
- "Which categories have more AI-search traffic?"
Input/Output
- Input modes:
text/plain - Output modes:
application/json,text/plain