How Morningstar built a financial advisor AI assistant powered by Amazon Bedrock AgentCore

This article describes how Morningstar built an AI assistant into its Direct Advisory Suite using Amazon Bedrock AgentCore to automate financial advisor workflows.

• Agentic orchestration decomposes advisor requests into multi-step workflows including portfolio analysis, research updates, and compliance checks
• AgentCore Runtime provides session isolation, VPC networking, custom JWT authorization, and durable memory for secure multi-tenant execution
• Amazon Bedrock Guardrails enforce policies on input and output to block prompt attacks, sensitive data exposure, and unauthorized advice patterns
• Advisor remains in control with explainable outputs, full audit trails, and human-in-the-loop recommendations rather than autonomous decisions
• Assistant grounds responses in Morningstar's proprietary research data, analytics, and workflow tools accessible via Model Context Protocol
• Observability through AgentCore and LangSmith enables tracing across multi-step plans, tool calls, and model invocations for production debugging

The AI assistant reduces advisor prep time from hours to minutes while maintaining regulatory compliance and data security for financial services.