Build custom code-based evaluators in Amazon Bedrock AgentCore
Amazon Bedrock AgentCore now lets you build custom evaluators—here's how to deploy them across financial agents in production.
Why it matters
AWS is expanding Bedrock's agent evaluation capabilities, allowing enterprises to deploy custom code-based validators at scale. This addresses a real production need: ensuring AI agents meet domain-specific safety and accuracy standards before runtime.
The key facts
11 to knowAmazon Bedrock AgentCore ships custom code-based evaluators
Four Lambda-based evaluator types demonstrated (financial market-intelligence use case)
Evaluators support both on-demand and online modes
Integration with AWS services for fact-checking, PII detection, real-time alerting
Composable with built-in evaluators
Amazon Bedrock AgentCore feature: custom code-based evaluators via Lambda
Use case: financial market-intelligence agents
Four evaluator types demonstrated: fact-checking, PII detection, real-time alerting, and custom validation
Integration with built-in evaluators (hybrid evaluation approach)
On-demand and online evaluation modes supported
Cross-service AWS integration (Bedrock + Lambda + other services for grounding)
Go to the source
AWS Machine Learning Blogaws.amazon.com
Publisher excerpt: In this post, you will implement four Lambda-based custom code evaluators for a financial market-intelligence agent, register each with AgentCore, and run them in on-demand and online modes. You will also see how to combine custom code-based evaluators with built-in evaluators and how to call other…