title: "AWS Rolls Out AgentCore Insights Preview, Expands Amazon Q Agents" slug: "aws-rolls-out-agentcore-insights-preview-expands-amazon-q-agents" published: "2026-07-12" beat: "Launches" tags: ["Launches", "Tools"] creator: "Agentry Newsroom" editor: "Susanne Sperling, Editor — Human in the Loop" tools: ["Claude (Anthropic)", "Perplexity Sonar"] creativeWorkStatus: "verified" dateReviewed: "2026-07-12" aiActArticle50: "compliant" humanView: "https://agentry.news/aws-rolls-out-agentcore-insights-preview-expands-amazon-q-agents" agentView: "https://agentry.news/agent/aws-rolls-out-agentcore-insights-preview-expands-amazon-q-agents"
AWS introduced new optimization capabilities for Amazon Bedrock AgentCore on June 17, 2026, at AWS Summit New York, enabling developers to analyze agent failures and improve performance in production.
Drafted by an AI agent. Verified by Susanne Sperling, Editor — Human in the Loop. AI policy.
AWS introduced optimization capabilities for Amazon Bedrock AgentCore on June 17, 2026, at AWS Summit New York, positioning the release as a step toward production-grade autonomous agent operations. Dr. Swami Sivasubramanian, AWS VP, highlighted the announcement in his keynote, detailing new tools to help teams understand and improve how their agents perform About Amazon.
The centerpiece is AgentCore Insights, a new diagnostic capability now in public preview that turns production traces into actionable intelligence. According to AWS documentation, Insights allows teams to "see failure, intent, and trajectory insights across hundreds of agent sessions, allowing you to understand what your agents are doing and ways to improve how they operate" AWS Bedrock AgentCore Guides. The tool discovers failure patterns, explains their root causes, and ranks improvement opportunities by impact—enabling continuous iteration on live agents without requiring manual audit cycles.
This feature addresses a known friction point in agent operations: production systems often fail silently or with incomplete logging. By systematizing trace analysis, AWS is making it easier for teams to move beyond rule-based guardrails and toward data-driven agent refinement.
Alongside Insights, AWS confirmed that AgentCore harness, recommendations, and A/B testing are now generally available AWS Bedrock AgentCore Guides. The company also announced GA status for Bedrock Managed Knowledge Base, Web Search integration, and guardrail integration, expanding the toolkit for building Amazon Q autonomous agents—the enterprise-focused agent product for background task execution.
Amazon Q agents can now retrieve company-specific data via managed knowledge bases and search the public web, while staying within safety constraints enforced by Bedrock Guardrails. This combination targets use cases where agents need both proprietary context and real-time information without human supervision.
The release reflects AWS's strategy to lower operational friction for agent deployments. By bundling diagnostics (Insights), safety (Guardrails), data access (Knowledge Base, Web Search), and optimization (A/B testing, recommendations) into a single service, AWS is reducing the engineering work required to move agents from prototype to production. The public preview status for Insights signals AWS expects rapid iteration based on early user feedback before full GA.
No pricing changes or performance benchmarks were announced. The features are available via the AWS Bedrock console and API today.