Transforming Productivity at New Relic: The Development of NOVA
Overview of New Relic and the Challenge
Introduction of New Relic NOVA: A New Era of AI
Solution Overview: Objectives and Architecture
Key Features of New Relic NOVA
Data Integration Strategy: Seamless Knowledge Retrieval
Leveraging Amazon Nova Models for Enhanced Performance
Advanced RAG Implementation for Optimized Retrieval
Evaluation Framework: Ensuring Accuracy and Relevance
Observability and Continuous Improvement Initiatives
Key Considerations and Future Directions
Conclusion: The Impact of NOVA on Organizational Efficiency
Further Reading and Author Insights
Transforming Productivity: The Innovative Journey of New Relic NOVA
In the realm of technology, New Relic Inc. stands out as a pioneer in application performance monitoring (APM) and comprehensive observability solutions. Headquartered in San Francisco, this leading tech company partners with esteemed brands worldwide to enhance digital system performance, ultimately delivering superior customer experiences. However, like many rapidly growing organizations, New Relic encountered a significant challenge: fragmented documentation and excessive time spent on internal system queries hindered productivity.
The Challenge
New Relic’s engineers found themselves overwhelmed, often spending days searching through disjointed informational silos. As an observability platform supporting thousands of customers, New Relic realized they needed a more efficient way to manage organizational knowledge. This recognition catalyzed the creation of New Relic NOVA (New Relic Omnipresence Virtual Assistant), an intuitive AI tool designed to revolutionize how New Relic employees access and utilize company information.
Enter New Relic NOVA
Developed using Amazon Web Services (AWS), New Relic NOVA evolved from a knowledge assistant to a full-fledged productivity engine, collaborating closely with the Generative AI Innovation Center. Harnessing AWS services like Amazon Bedrock and Amazon Kendra, New Relic NOVA transformed internal workflows, generating intelligent code reviews, AI governance, and managed Model Context Protocol (MCP) services.
Amazon Bedrock allows the use of various foundational AI models, facilitating easy customization and ensuring that teams can adapt the models to specific operational needs—without the hassle of infrastructure management.
Solution Architecture
New Relic NOVA was built with a clear focus on enhancing productivity while ensuring data security and consistent response quality. The framework underlying NOVA integrates various AWS services, maintaining an agile approach that allows the solution to evolve with the organization’s expanding needs.
Key architectural components include:
- Main Agent Layer: Acts as an orchestrator, identifying user intent and routing requests to specialized downstream layers.
- Data Source Layers: Ingests and enriches internal knowledge, enhancing retrieval performance and relevancy.
- Agents Layer: Comprises both Strands Agents for third-party services and custom action agents for internal tasks.
Data Integration and Advanced Capabilities
New Relic NOVA showcases exceptional data integration with multiple sources, maximizing accuracy and rapid access to crucial information. By employing technologies like Amazon Kendra for GitHub Enterprise and enhancing Slack channel history with Amazon Q Index, New Relic NOVA streamlines workflows and ensures prompt access to relevant data.
Moreover, the deployment of Amazon Nova models within the AWS framework optimizes response times while balancing performance and cost, providing a seamless experience for users.
Continuous Improvement and Feedback Mechanisms
New Relic NOVA emphasizes the importance of a robust feedback loop. By continually collecting user insights through unique channels like Slack reactions, New Relic can fine-tune the performance of NOVA. Metrics such as answer accuracy and context relevance are tracked, ensuring perpetual enhancements aligned with real user interactions.
Lessons Learned and Future Directions
- Understanding User Pain Points: Engaging with end-users from the project’s inception ensures that solutions are tailored effectively.
- Data Integration Strategies: Implementing robust strategies that enhance data relevance is pivotal.
- Feedback Mechanisms: Real-time feedback loops facilitate ongoing improvements and user adoption.
In the near future, New Relic anticipates the evolution of NOVA into a comprehensive enterprise AI platform. By leveraging advanced AWS technologies, the team aims to explore cost-efficient solutions like Amazon S3 Vectors, facilitating the handling of large-scale AI workloads.
Conclusion
The journey of New Relic NOVA exemplifies how leveraging generative AI technologies can fundamentally reshape organizational productivity and operational effectiveness. By integrating services like Amazon Bedrock and Kendra, New Relic significantly reduced information retrieval times while automating complex tasks.
To explore how transformational generative AI can enhance your business operations, consider connecting with the AWS Generative AI Innovation Center or an AWS Partner Specialist. Embrace the future—like New Relic, you too can unlock the full potential of your organizational knowledge.
Further Reading and Resources
If you’re interested in the technicalities behind New Relic NOVA or wish to understand how AWS services can fit into your business needs, feel free to dive deeper into the resources available through AWS and New Relic.
About the Authors
- Yicheng Shen: Lead Software Engineer at New Relic NOVA focused on building intelligent systems.
- Sarathy Varadarajan: Senior Director of Engineering at New Relic, driving AI-first transformation.
- Joe King: Senior Data Scientist at AWS’s Generative AI Innovation Center, architecting generative AI solutions.
- Priyashree Roy: AWS Data Scientist applying machine learning to build innovative solutions.
- Gene Su and Dipanshu Jain: Specialists at AWS in generative AI, enhancing various sectors through technology.
- Ameer Hakme: AWS Solutions Architect aiding customers in harnessing AI capabilities effectively.