German insurance broker MRH Trowe deployed self-service AI agents for 400 employees on the AWS platform
German insurance broker MRH Trowe made self-service AI agents available to 400 employees using Strands Agents, Amazon Bedrock AgentCore and LibreChat, according to the AWS blog, with costs of around 14 USD per employee per month and data in the Frankfurt region.
MRH Trowe, a German insurance broker operating mainly in Germany, Switzerland and Austria, deployed self-service AI agents for approximately 400 employees, according to an article published on the AWS Machine Learning blog. The deployment reached this scale during the first month of production operation. The solution combines the open source SDK Strands Agents for building agents, the Amazon Bedrock AgentCore platform for running and scaling them, and the open source chat interface LibreChat as the user interface for employees.
According to the company, initial operating costs were approximately 14 dollars per employee for the first month, with the potential to reduce infrastructure costs by roughly 40 percent through resource rightsizing and scheduled scaling. The aim was to enable employees to create and use AI agents without deep technical knowledge, while remaining within a centrally managed environment that meets the data protection and compliance requirements of the German financial sector. Another motivation was that individual teams had already begun experimenting with AI independently, which, according to the company, created a risk of fragmented, unmanaged tools and potential leaks of sensitive client data.
The first agent deployed to production is a tool that turns a Microsoft Teams meeting transcript into structured minutes with the date, participants, agenda and action items – an employee requests it in German directly in LibreChat. Each request runs under the identity of the signed-in employee, authenticated through Microsoft Entra ID, so the agent has access only to that employee's own calendar and transcripts, and all data processing takes place in the AWS Europe (Frankfurt) region, in Germany.
The remainder of the architecture description (network configuration, security layers and other details) is not available in full in the sources. Details can be found in the source article.
Why it matters
The case demonstrates a concrete solution to a dilemma faced by many regulated companies: how to enable employees to use generative AI widely without the risk of sensitive data leaks or the emergence of unmanaged “shadow” tools. The documented costs (14 USD per user per month) and an architecture built on a combination of open source and managed services give other companies in regulated industries a measurable reference point for their own deployment decisions.
Two audiences, two different impacts
What this means
For individuals
For developers and IT architects, this is a concrete reference combination of open source and managed tools (an agent SDK, an isolated environment for running agents, a chat interface connected to corporate identity) that can serve as a model when designing a similar solution.
For a business
Companies in regulated industries (finance, insurance) gain a documented example of how to make generative AI available to a large number of employees without losing control over data and without fragmented, unapproved tools, with measurable costs per user.
ProcessesCheck the original
Event sources
only one source so far · 1 publisher, 0 independent. We count feeds from the same owner only once.