AWS AI Updates: May 29, 2026

1. Claude Opus 4.8 Arrives on Bedrock and Claude Platform on AWS

AWS. Anthropic’s Claude Opus 4.8 is now available through both Amazon Bedrock and the Claude Platform on AWS, giving teams a choice between Bedrock’s managed Guardrails and Knowledge Bases or the native Anthropic console experience. AWS positions the model for longer autonomous runs, deeper reasoning, and production coding workflows, citing its ability to read codebases, plan before editing, and hold context across long sessions. The Bedrock path keeps inference data inside AWS infrastructure, which matters for regulated deployments. Source

2. Next-Generation OpenSearch Serverless Targets Agentic and Vector Workloads

AWS. The next generation of Amazon OpenSearch Serverless is generally available as a fully managed search and vector engine built for agent workloads, auto-scaling roughly 20x faster than its predecessor and provisioning resources in seconds. It fully decouples compute and storage with scale-to-zero and pay-per-usage pricing, which AWS says can cut costs up to 60 percent versus provisioning for peak. The release adds integrations with AI development platforms including Vercel and Kiro, plus OpenSearch Agent Skills for coding platforms. Source

3. AWS Resilience Hub Adds Generative AI Failure Mode Analysis

AWS. A new generation of AWS Resilience Hub introduces a generative AI-powered failure mode assessment that evaluates services against AWS Well-Architected best practices and produces prioritized recommendations using the AWS Resilience Analysis Framework. Dependency assessments surface the AWS services, internal endpoints, and third-party services an application relies on, and a new three-level hierarchy of systems, user journeys, and services supports organization-wide reporting through AWS Organizations. The service is live in 15 regions with self-paced migration for existing customers. Source

4. Amazon Connect Extends Generative AI Post-Contact Summaries to Eight More Languages

AWS. Amazon Connect now generates generative AI post-contact summaries in eight additional languages, including Portuguese, French, Italian, German, Spanish, Chinese, Japanese, and Korean, along with British and Australian English variants. Summaries are produced in the conversation’s original language across voice, chat, and email, letting agents and supervisors skip full transcripts while monitoring quality across multilingual teams. The languages are available in all regions where Connect post-contact summaries already run. Source

5. Amazon Connect Uses Generative AI to Evaluate Self-Service Interactions

AWS. Amazon Connect can now apply generative AI to automatically evaluate self-service interactions against custom natural-language criteria, returning detailed reasoning for each quality assessment rather than a bare score. This extends AI-driven evaluation beyond agent-handled contacts to bot and IVR flows, where measuring containment quality has historically required manual sampling. The capability lets contact center teams audit automated experiences at scale with explainable results. Source