Amazon Web Services

Latest AI news, models and releases from Amazon Web Services. ['AgentCore', 'Amazon API Gateway', 'Amazon Bedrock', 'Amazon Bedrock AgentCore', 'Amazon Bedrock Guardrails', 'Amazon Bedrock Knowledge Bases', 'Amazon Bedrock Managed Knowledge Base', 'Amazon DynamoDB', 'Amazon Elastic Kubernetes Service', 'Amazon GuardDuty', 'Amazon Nova 2', 'Amazon Nova 2 Lite', 'Amazon Nova Act', 'Amazon Nova Pro', 'Amazon Quick', 'Amazon Quick Flows', 'Amazon Quick Research', 'Amazon Quick Sight', 'Amazon SageMaker AI', 'Amazon Simple Storage Service', 'Amazon Textract', 'Amazon Titan', 'Amazon Titan Embeddings', 'Amazon Titan Text', 'Amazon Titan Text Embeddings V2', 'AWS AI Security Framework', 'AWS Lambda', 'AWS Step Functions', 'Bedrock', 'Bedrock Managed Knowledge Base', 'Claude 3.7 Sonnet', 'Claude 4.5 Sonnet', 'Claude 4 Sonnet', 'NEXUS', 'Nova Act', 'Strands Agents']

Agents 🇺🇸

Evaluating AI Agents: A production blueprint with Strands and AgentCore from AWS and Motorway

Motorway, a UK online car marketplace, partnered with AWS to build an AI-powered dealer stock search agent. They developed an evaluation pipeline combining Strands Agents SDK and Amazon Bedrock AgentCore, reducing incorrect results from 1 in 8 queries to 1 in 50 and cutting issue detection time from hours to minutes. The blueprint includes a three-layer evaluation framework for tool usage, reasoning, and output quality, with quality gates blocking deployments when metrics fall below thresholds.

Amazon/AWSAmazon/AWS Amazon Web ServicesAmazon Web Services AnthropicAnthropic
AWS ML blog23.07 · 23:49
Fresh news