{
  "id": 5148327,
  "title": "How an AWS team detects dashboard content failures at scale using Amazon Bedrock",
  "url": "https://urgent.news/2026/09/02/how-an-aws-team-detects-dashboard-content-failures-at-scale-using",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-02T18:21:20.000Z",
  "source": {
    "name": "AWS Machine Learning",
    "slug": "aws-machine-learning",
    "url": "https://aws.amazon.com/blogs/machine-learning/how-an-aws-team-detects-dashboard-content-failures-at-scale-using-amazon-bedrock/"
  },
  "original_language": "en",
  "account": "A large share of organizations maintain business intelligence (BI) systems on an immense scale. Imagine a situation where an executive prepares for an important meeting and encounters an empty chart on their dashboard. All the infrastructure monitors are functioning correctly. The servers are up, APIs are responding, and the data pipeline completed without any issues. However, the content displayed on the screen is problematic, and the system does not flag this issue. This type of failure is subtle—it only manifests in what the user sees. Consequently, it goes undetected by infrastructure monitoring systems and requires users to report it manually. Our subsequent investigations revealed that this type of failure occurs in under 1 percent of cases. Dashboard components, such as tables, charts, and visuals, can exhibit blank, outdated, or inaccurate data. Potential causes include upstream pipeline failures, alterations in permissions, and temporary infrastructure issues. Even if all charts appear correctly, the numbers displayed may be erroneous. This risk escalates as businesses integrate dashboard data with AI systems that generate business narratives. In this post, we will detail the creation of a last-mile automated content validation solution. This solution scans hundreds of dashboards hosted on the AWS Insights application (powered by Amazon Quick ) to identify missing or incorrect elements. With this solution, the BI and analytics team can identify issues before users notice them. The solution monitors dashboards continuously and analyzes visuals using large language models (LLMs) on Amazon Bedrock . The system then sends real-time alerts to developers when a health problem is discovered, reducing mean time to detection from up to 72 hours to less than 1 hour. This article covers the content-layer monitoring gap and why user reports alone are insufficient in detecting content failures. It also explains a five-stage serverless validation architecture built on AWS managed services. Additionally, we will explore two parallel AI validation mechanisms, one targeting visual integrity and the other focusing on numeric consistency, built on the same design principles. The article concludes with production engineering lessons on managing false positives and maintaining LLMs away from arithmetic.",
  "summary": "Business intelligence dashboards can fail silently, showing blank, stale, or wrong data even when every infrastructure monitor reports healthy. Learn how an AWS team built an automated, AI-powered content validation solution on Amazon Bedrock that scans hundreds of dashboards and alerts owners, cutting mean time to detection from days to under an hour.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}