Gen Digital's postmortem: building Norton Revamp's AI news engine on AgentCore Web Search
AWS ML Blog · rss · 2026-11-04
Gen Digital's engineering team recounts how they built the AI-curated News Engine at the heart of Norton Revamp, a generative AI professional reputation management platform launched in March 2026, using Amazon Bedrock AgentCore Web Search.
Why AgentCore Web Search: they evaluated custom crawlers (heavy maintenance, stale), third-party news APIs (data leaving AWS, vendor complexity), and static knowledge bases (perpetually stale). AgentCore offered fully managed search, zero data egress, and cited sources with URLs and dates.
Three features: a goal-aware personalized news feed, same-day trending content suggestions in the user's authentic voice (Analyze→Refine→Post flow for LinkedIn and X), and trend-aware weekly goal roadmaps.
Stack: Amazon Nova Lite for generation, Claude for complex reasoning, Stability AI for images; event-driven architecture with EventBridge daily scans, Bedrock batch analysis of up to 1,000 historical posts, Terraform with blue-green deploys.
AI-native development results: AI generated 70% of UI code; design-to-prototype went from weeks to hours; 3x faster features, 90% faster boilerplate, 80% docs, 70% tests—offset by 20-30% of time spent reviewing AI output.
More from coding & agent
- Monologue's Playbook: When to Adapt Your Own AI Model — every · 2026-11-06
- Building a Physical Codex Pet with Astra and Some CAD — OpenAIDevs · 2026-11-06
- SkipLabs joins Mistral to build language tools for AI-written code — syhw · 2026-11-06
- LukeW's Intent redesigns UI around workspaces, not chat threads — LukeW · 2026-11-06
- Start evals from 100 production traces: a 7-step recipe from Hamel Husain's FAQ — HamelHusain · 2026-11-06
- Dev releases pixel art model for game dev with exact size control and sprite-based style consistency — HugoDuprez · 2026-11-06