Skip to main content

Creating a Knowledge Graph

Knowledge Graphs in Amplifi transform your unstructured data into a connected network of entities and relationships, enabling deeper insights and more intelligent search capabilities.

Once your dataset is ingested and processed, you can generate a Knowledge Graph to automatically extract entities (people, organizations, concepts, locations) and discover how they relate to each other within your content.

What You'll Get​

A Knowledge Graph provides:

  • Entity Recognition: Automatically identifies key entities from your documents
  • Relationship Mapping: Discovers connections between entities across your data
  • Enhanced Search: Powers relationship-aware search that goes beyond keyword matching

Step 1: Start Graph Creation​

After your dataset is fully ingested, navigate to the dataset details page and click + Add Graph to Dataset to initiate graph generation. Amplifi will automatically analyze your content using advanced NLP models to identify entities and their relationships.

You can choose between:

  • Extract from Dataset: Let Amplifi automatically detect relevant entities using AI-powered entity recognition
  • Define Manually: Manually specify entities you want to include for more targeted graph creation

Step 2: Review and Select Entities​

Once entities are extracted, Amplifi displays a comprehensive list of detected entity types (e.g., Person, Organization, Location, Product, Date). Review the detected entities and select which types to include in your Knowledge Graph.

This step allows you to:

  • Filter out irrelevant entity types
  • Focus on entities most important to your use case
  • Ensure the graph contains the right level of detail

Step 3: Generate the Graph​

After confirming your entity selection, click Generate Graph to create your Knowledge Graph. Amplifi will process your dataset in the background, analyzing relationships between entities and building a structured network.

The generation process typically takes a few minutes depending on your dataset size. You'll receive a notification when the graph is ready.

Step 4: View Graph Summary​

Once generation is complete, Amplifi displays a comprehensive summary showing:

  • Entity Types: Categories of entities found (Person, Organization, Location, etc.)
  • Entity Counts: Number of entities identified in each category
  • Relationship Count: Total connections discovered between entities

This summary helps you understand the scope and quality of your Knowledge Graph. You can view:

  • Entity Type Distribution: See which types of entities are most prevalent in your data
  • Individual Entities: Browse specific entities within each category to understand your data composition
  • Entity Relationships: Examine how entities are connected and the strength of their relationships

Step 5: Explore Relationships​

Navigate to the Graph Relationships tab to explore the discovered connections. This interactive view shows:

  • Entity Connections: How different entities relate to each other
  • Relationship Types: The nature of connections (works_with, located_in, manages, etc.)

Step 6: Manage the Graph​

Once your Knowledge Graph is created, you can:

  • Regenerate Graph: Update the graph when your dataset changes or you want to modify entity selection
  • Edit Entity Selection: Add or remove entity types to refine the graph
  • Delete Graph: Remove the graph if it's no longer needed

Best Practices​

  • Start Small: Begin with a focused dataset to understand graph quality before scaling up
  • Review Entity Types: Carefully select relevant entity types to avoid noise in your graph
  • Iterate and Refine: Regenerate graphs as your data evolves to maintain accuracy

Benefits of Knowledge Graphs​

Creating a Knowledge Graph unlocks powerful capabilities:

  • Smarter Search: Agents can leverage entity relationships for more contextual answers
  • Pattern Discovery: Identify hidden connections and patterns across your data
  • Enhanced Analytics: Build more sophisticated queries using entity relationships
  • Better Insights: Understand how concepts, people, and organizations interact in your content

Knowledge Graphs transform your data from isolated documents into a connected, intelligent network that amplifies the power of AI-driven analysis.