Gen Digital Uses Amazon Bedrock AgentCore Web Search to Power Norton Revamp News Engine
AI assistants · AI coding · Enterprise AI
Gen Digital integrated Amazon Bedrock AgentCore Web Search to ground Norton Revamp's AI news feed and content suggestions with real-time web data. The team reported maintaining AWS data boundaries while using AI tools across development.
Gen Digital developed an AI-curated News Engine for Norton Revamp using Amazon Bedrock AgentCore Web Search. According to the team, the platform grounds content recommendations, goals, and feeds in cited web data within the AWS environment to avoid external data egress. The technical stack combines Amazon Bedrock, Amazon Nova Lite for text generation, Stability AI for visuals, and Claude for complex reasoning. The team stated that using AI during engineering accelerated feature delivery and boilerplate code generation, while human review accounted for 20 to 30 percent of development time.
Key takeaways
- Norton Revamp uses AgentCore Web Search via MCP to retrieve cited web snippets and publication dates directly within AWS for real-time grounding.
- The architecture combines Amazon Nova Lite for query and content generation, Bedrock batch analysis, and an internal ranking system.
- The engineering team reported generating 70 percent of UI code with AI assistance, while human engineers handled architecture, security, and algorithm design.
- Testing achieved 85 percent code coverage, with the team noting 99.9 percent uptime since launch alongside continuous content monitoring.
Source：AWS Machine Learning Read the original ↗
Prepared, checked and automatically published by AI; not individually reviewed by a human.