What are Related Prompts
Related prompts are actionable recommendations surfaced within the Empathy Platform. They guide people to explore, filter, or engage with content and products in a contextually relevant way.
A related prompt consists of:
- Suggestion text: A brief description of the recommendation.
- Optional image: An illustration that visually supports the recommendation.
- Actionable content: The core of the prompt, which can be:
- Expansive prompts: A list of search queries to broaden exploration.
- Item recommendations: A list of products or items relevant to the user’s context.
- Reductive prompts: A set of filters to narrow down search results.
- Conversational chat: An introductory text to open a chat-based interaction.
interact
Willing to know more about Related Prompts? Check the Related Prompts overview page to keep reading.
Structure of a related prompt
Each related prompt follows a standardized schema, making it easy to parse and act upon. Here is the general structure:
{
"suggestionText": "Description of the recommendation",
"image": {
"url": "https://example.com/image.jpg",
"altText": "Image description"
},
"metadata": {
"origin": "source/origin",
"createdAt": "ISO timestamp",
"model": "model-name",
"source": "optional-source-id"
},
"context": {
"lang": "language-code",
"instance": "platform-instance",
"vertical": "vertical",
"categories": ["category1", "category2"],
"subcategories": ["subcategory1", "subcategory2"]
},
"id": "unique-id",
"type": "prompt-type",
"content": {
// Varies by type (see examples below)
}
}
These are some examples of related prompts based on the type of actional content:
Expansive recommendation
This example shows how to use related prompts to broaden the shopper's exploration with suggested search queries:
{
"suggestionText": "Explore these related topics about machine learning",
"image": {
"url": "https://example.com/images/machine-learning-topics.jpg",
"altText": "Neural network visualization"
},
"metadata": {
"origin": "online",
"createdAt": "2023-06-15T08:30:00Z",
"generation_model": "deepseek/deepseek-R1",
"source": null
},
"context": {
"lang": "en",
"instance": "mymotivemarketplace",
"vertical": "art-and-books",
"categories":["Books"],
"subcategories": ["Machine Learning", "Technology"]
},
"id": "rec-ml-topics-123",
"type": "expansive",
"content": {
"searchQueries": [
"machine learning basics",
"neural networks tutorial",
"deep learning frameworks comparison",
"AI ethics concerns",
"machine learning job market"
]
}
}
Item recommendation
This example shows how to use related prompts to recommend specific products or items.
{
"suggestionText": "Check out these camping essentials for your next trip",
"image": {
"url": "https://example.com/images/camping-gear.jpg",
"altText": "Camping equipment laid out on grass"
},
"metadata": {
"origin": "curated",
"createdAt": "2023-07-22T14:15:30Z",
"model": "deepseek/deepseek-R1",
"source": "outdoor_expert_123"
},
"context": {
"lang": "en",
"instance": "mymotivemarketplace",
"vertical": "art-and-books",
"categories":["Books"],
"subcategories": ["Machine Learning", "Technology"]
},
"id": "rec-camping-456",
"type": "item-recommendation",
"content": {
"products": [
{
"id": "prod-789",
"name": "3-Season Backpacking Tent",
"description": "Lightweight, waterproof tent perfect for spring through fall camping",
"imageUrl": "https://example.com/images/products/tent.jpg"
},
{
"id": "prod-790",
"name": "Portable Camping Stove",
"description": "Compact butane stove that folds for easy packing",
"imageUrl": "https://example.com/images/products/stove.jpg"
},
{
"id": "prod-791",
"name": "Water Purification System",
"description": "Filter that removes 99.9% of bacteria from stream water",
"imageUrl": "https://example.com/images/products/water-filter.jpg"
}
]
}
}
Reductive recommendation (search filters)
This example shows how to use related prompts to help people narrow down search results using filters.
{
"suggestionText": "Narrow your laptop search with these specifications",
"image": null,
"metadata": {
"origin": "online",
"createdAt": "2023-08-05T10:22:45Z",
"model": "deepseek/deepseek-R1",
"source": null
},
"context": {
"lang": "en",
"instance": "mymotivemarketplace",
"vertical": "art-and-books",
"categories":["Books"],
"subcategories": ["Machine Learning", "Technology"]
},
"id": "rec-laptop-filter-789",
"type": "reductive",
"content": {
"filters": [
{
"name": "processor",
"value": ["Intel i7", "AMD Ryzen 7"]
},
{
"name": "ram",
"value": "16GB"
},
{
"name": "ssd",
"value": "512GB"
},
{
"name": "price_range",
"value": {
"min": 800,
"max": 1500
}
},
{
"name": "has_discrete_gpu",
"value": true
}
]
}
}
Conversational chat
This example shows how to use related prompts to invite people to start a chat-based interaction.
{
"suggestionText": "Are you looking for a recipe? Let's chat!",
"image": {
"url": "https://example.com/images/chef-assistant.jpg",
"altText": "Chef in kitchen preparing food"
},
"context": {
"lang": "en",
"instance": "mymotivemarketplace",
"vertical": "art-and-books",
"categories":["Books"],
"subcategories": ["Machine Learning", "Technology"]
},
"metadata": {
"origin": "offline",
"createdAt": "2023-09-12T16:30:00Z",
"model": "deepseek/deepseek-R1",
"source": "recipe_assistant_v2"
},
"id": "rec-recipe-chat-101",
"type": "conversational-chat"
}