MRH Trowe Rolls Out Secure Self-Service AI Agents to 400 Employees Using AWS Bedrock

MRH Trowe, a commercial and industrial insurance broker based primarily in Germany, provided approximately 400 employees with secure access to self-service AI agents during its first month of production. [1] The company utilized a framework combining Strands Agents, Amazon Bedrock AgentCore, and LibreChat to meet the data residency and compliance requirements of the German financial sector. [1] Initial production costs averaged around $14 per seat in the first month, and the firm identified a pathway to reduce infrastructure expenses by approximately 40 percent through scheduled scaling and right-sizing. [1]

This deployment demonstrates how heavily regulated organizations can mitigate the risks of unmanaged AI experimentation while empowering employees to build agents without deep technical expertise. By combining AWS Bedrock AgentCore with Strands Agents and LibreChat, MRH Trowe established a centrally governed environment that prioritizes data residency and cost transparency over ad-hoc shadow IT tools. What the sources don't address: How the company tracks the specific productivity gains or revenue generated by these 400 employees to justify the ongoing per-seat licensing and infrastructure costs.