Business application candidate agent
Summarize
Summary of Business application candidate agent
The Business application candidate agent in ServiceNow automates the discovery and suggestion of business applications linked to existing application services in the CMDB. It leverages generative AI clustering and continuous feedback loops to propose candidate business applications, helping reduce manual mapping efforts and enhance CMDB data governance and quality. Users validate AI-generated recommendations by accepting or rejecting them, which further refines recommendation accuracy.
Show less
Key Features
- AI-driven grouping: Uses generative AI to cluster related application services and infer the business applications they represent.
- Feedback integration: Incorporates user acceptance and rejection feedback to improve future recommendations.
- Automated data synchronization: Runs scheduled background jobs to sync application service and business application data into AI-optimized tables for efficient processing.
- Recommendation management: Stores candidate business application suggestions in a dedicated table for human review and decision-making.
- Governance control: Requires the sncmdbadmin role for access to agent tables, forms, and configurations, ensuring controlled management of the discovery process.
How It Works
The agent operates via two synchronized background jobs:
- Application Service Data Synchronization: Copies and updates application services and business applications into internal tables optimized for AI processing. This job runs weekly by default or more frequently during initial setup.
- Recommendation Processing: Generates business application candidates based on AI clustering of unsynchronized app services, incorporating user feedback from prior cycles to enhance accuracy.
The system includes scheduled jobs to activate or deactivate the agent, work queues to monitor processing stages, and tables to maintain configuration, recommendation records, and statistics.
Practical Benefits for ServiceNow Customers
- Reduced manual effort: Automates the complex task of mapping numerous application services to business applications.
- Improved CMDB quality: Enhances data consistency and accuracy by discovering and suggesting relationships that may be overlooked manually.
- Adaptive learning: Continuously improves recommendation quality through customer feedback on accepted or rejected suggestions.
- Governed AI use: Encourages validation of AI-generated data to maintain organizational governance standards.
Next Steps for Customers
Administrators with the sncmdbadmin role should review and evaluate the agent’s business application candidate recommendations regularly, accepting valid mappings to create new business application records and rejecting inaccurate suggestions to refine AI learning. This process helps maintain an up-to-date and well-governed CMDB that accurately reflects business service relationships.
The Business application candidate agent discovers and suggests business applications to associate with existing application services in the CMDB, reducing manual mapping effort and improving data governance. The agent uses AI clustering and feedback loops to propose business application candidates for your review.
Generative AI might produce inaccurate or incomplete information. Always validate AI-generated recommendations and accept or reject them based on your organization's governance policies.
Required role
Access to all business application candidate agent tables, forms, and configuration requires the sn_cmdb_admin role.
What is it
The agent uses existing service instance and user group data to infer business applications and to generate candidate records that admins can review and promote to actual records.
Business applications are logical representation of business functions. It is challenging to manually map hundreds of application services to business applications. The Business application candidate agent uses generative AI to automatically group related app services, infer the business applications they collectively represent, and surface these as reviewable recommendations. Each recommendation can be accepted or rejected — and rejections feed back into the AI to improve future results.
The agent surfaces application services that have no linked business application and proposes new business application records for your review and approval. The agent uses generative AI clustering, ServiceNow Otto skills, and continuous feedback from your rejection decisions to improve recommendation quality over time.
Key benefits
The Business application candidate agent provides the following benefits:
- Reduces manual effort required to map application services to business applications
- Improves CMDB data quality and consistency through automated relationship discovery
- Uses AI to identify patterns and relationships humans might miss
- Learns and improves from human feedback on rejected recommendations
How it works
The Business application candidate agent runs as two synchronized background jobs that manage data and generate recommendations:
The job copies app services and business applications into internal, AI-search-optimized tables where the AI engine can access them efficiently.
- The job runs weekly by default, or more frequently during initial setup.
- On the first run, the job bulk-loads all records; subsequent runs pick up only changed records.
- The job completes before the Processing job begins generating recommendations.
The AI captures human acceptance and rejection decisions. Uses rejection decisions to refine its clustering and matching logic, improving future recommendations.
System overview
The Business application candidate agent operates through the following system components:
- Scheduled Jobs: Control whether the agent is active or deactivated.
- Work queues: Track processing progress through the synchronization and recommendation stages.
- Run-state table: Maintains statistics and configuration for each processing cycle.
- Recommendation table: Stores recommendations for human review, acceptance, or rejection.