Using Enterprise graph schema
Summarize
Summary of Using Enterprise Graph Schema
Enterprise Graph is a pre-configured Knowledge Graph schema designed to enhance natural language query responses across the entire ServiceNow database. This schema maps all instance tables and their connections, simplifying the process of Knowledge Graph setup and allowing administrators to utilize AI agents and the Now Assist Virtual Agent effectively.
Show less
Key Features
- Comprehensive Query Capability: Enterprise Graph expands query capabilities to include all tables in the database, unlike custom schemas that are limited to specific tables.
- Integration with AI Agents: The schema allows AI agents and Now Assist tools to access structured data efficiently, improving response accuracy and context for tasks.
- Two Operational Modes:
- Regular Mode: Suitable for use cases with over 50 tables, prioritizing responses based on tagged tables.
- Small Mode: Limits searches to tagged tables only, ideal for latency-sensitive scenarios with fewer than 50 tables.
- Tagging Requirement: Tags are essential for both modes, defining key tables for natural language queries and ensuring relevant answers.
Key Outcomes
By implementing the Enterprise Graph schema, ServiceNow customers can expect improved efficiency and accuracy in data retrieval through natural language queries, enhanced productivity for fulfillers using the Now Assist panel, and effective query deflection from the Now Assist Virtual Agent. This integration supports a broader range of enterprise data inquiries and fosters a streamlined interaction with the ServiceNow platform.
Use Enterprise Graph for accurate natural language query responses, across the entire database.
Enterprise graph is a pre-configured Knowledge Graph schema that maps all instance tables and their connections, enabling natural language queries for data across all tables.
To see some examples of the Natural Language query responses, refer Natural language queries use cases and examples.
Enterprise graph schema simplifies Knowledge Graph setup by providing a preconfigured schema, eliminating the need for custom schema creation in Knowledge Graph designer.
Admins can choose Enterprise graph as the Knowledge Graph schema when using AI agents, Now Assist Virtual Agent, or Now Assist panel admin and add tags to enhance accuracy.
By mapping all tables, the Enterprise graph schema expands query capabilities to cover the entire database, whereas a custom or out-of-the-box (OOTB) schema limits queries to only the tables included in its specific schema.
Benefits of Enterprise Graph on Prebuilt integration
Now Assist Virtual Agent- With Enterprise Graph integration, Now Assist Virtual Agent can respond to a wide range of questions about their enterprise data from requester, helping to deflect queries.
Now Assist panel- With Enterprise Graph enabled, Now Assist panel allows fulfillers to obtain insights from the structured data in instance by asking natural language questions, which boosts their productivity.
AI agents- With Enterprise Graph enabled, AI agents can access relevant information from structured data in ServiceNow instance directly from the Knowledge Graph, giving them essential context for tasks.
Modes of Enterprise Graph
- Enterprise Graph- regular mode
- Enterprise Graph (Small)
Regular mode is used for use cases with large number of tables (more than 50 tables), to:
- Searches across all tables, but gives more priority to tables which are included in the tag added at time of configuring Enterprise Graph in the consuming app (AI agent or VA).
- Useful for scenarios where answer is expected from large number of tables.
- Searches only within tables which are included in the tag added at time of configuring Enterprise Graph (Small) in consuming app (AI agent or VA). Tagging is mandatory in this mode.
- Useful for latency-sensitive use cases with limit of 50 tables
Tags in Knowledge Graph
Tags are lists of key tables that are important for answering natural language questions in a specific use case. They are required to be used with Enterprise Graph and Enterprise Graph (Small).