# Simularity and LLMs

# Technical Support: AI Use Cases (Example)

This page collects AI patterns shown in the demo repository, rewritten as reusable building blocks for Curiosity Workspace applications.

# Use case: “suggest similar items” with constraints

A robust similarity experience is rarely “pure semantic search”. Instead:

• retrieve semantic neighbors (embeddings)
• constrain to a context (graph relationship, facet, or ownership boundary)
• return results with scores so the UI can explain ordering

See the product-first patterns:

• Semantic Similarity
• Embeddings

# Use case: AI-assisted chat with custom orchestration

The demo repository illustrates a common pattern:

• the UI uses a chat component
• message submission is routed through a custom endpoint
• the endpoint can inject domain context and trigger generation

This pattern is useful when you need:

• custom prompts and context policies
• tool availability control
• auditability and safe side effects

See:

• Prompting Patterns
• Custom Endpoints