Rede Mater Dei deployed 12 AI agents on Amazon Bedrock AgentCore and achieved 517% ROI

Brazilian hospital network Rede Mater Dei shared how it uses Amazon Bedrock AgentCore to monitor and manage AI agents in the revenue cycle. The company has already launched the first agents from a set of 12 and claims that over four months it achieved 517% ROI, reducing authorization times and accelerating the start of operations.

Why hospitals automate

The main problem Rede Mater Dei is trying to solve is the growth of denials on insurance and medical claims in Brazil. According to data presented in the article, in 2024 the average denial rate in the country's private hospitals grew from 11.89% to 15.

89%, and lost revenue across the market could reach 10 billion Brazilian reals. For a large hospital network, this is not just an accounting metric: every delay in contract verification, authorizations, and billing hits service timelines, team workload, and cash flow. Within the processes themselves, the network faced typical industry bottlenecks: lots of manual work, scattered documents, high turnover in operations teams, and constant re-checks at vulnerable stages.

In such a loop, AI agents are interesting not only as a way to save time, but also as a tool for standardizing solutions. If an agent participates in a critical stage of the revenue cycle, the company can no longer simply launch a model—it needs to see exactly what the agent did, why it did it, and how stably the entire system works.

How the system works

Rede Mater Dei, together with A3Data and AWS, built a program of 12 AI agents that should cover the entire hospital revenue cycle—from working with contract rules to authorizations and billing. Amazon Bedrock AgentCore became the base platform, providing a serverless environment for running agents, tool integration, memory, and observability capabilities. In the article, this set is called a digital workforce: agents should perceive data, make decisions, and perform actions as autonomously as possible, but in a managed and verifiable architecture.

• Contracts Agent collects and structures complex contract rules that were previously scattered across different documents.
• Parameterization Agent automatically transfers these rules into the hospital's ERP system and reduces the risk of manual errors.
• Authorization Agent handles requests, verification, and interaction with insurance companies.
• AgentCore Observability and AgentCore Evaluations add a monitoring layer where you can measure correctness, accuracy, usefulness, safety, and context relevance.

The architecture was divided into three layers. DEL is responsible for data preparation and placement into a structured data lake, AEL handles agent orchestration and execution, and TCL is responsible for trust, compliance, and traceability. Particular emphasis was placed on evaluating agent performance: Rede Mater Dei uses AgentCore Evaluations as a continuous improvement loop, where the quality of a multi-agent system can be measured by clear metrics rather than subjective team impressions.

What results already exist

The most notable part of the case is the numbers. According to project data, the first phase, focused on Observability and Evaluations, delivered the network 517% return on investment over the first four months. In parallel, authorization time was reduced by 66%, and time to start operations was reduced by 33%. For a medical network this is an important point: it's not just about internal efficiency, but also about more predictable procedure flow, which directly impacts scheduling, revenue, and patient experience.

"Our goal is to increase accuracy, predictability, and speed at critical stages of the revenue cycle," says

Rede Mater Dei Vice President Renata Salvador Grande.

The second effect is manageability. Thanks to unified telemetry, the team gets a complete audit trail of key decisions: which rules were applied, what actions the agent performed, and where failures or anomalies occurred. This is especially important in sensitive areas like contracts, authorizations, and billing, where any error can lead to payment denial or regulatory risks. Plus management gets KPIs in real time: volume of automated checks, assessment of financial impact, processing speed, and denial risk by insurers.

What this means

The Rede Mater Dei case shows that the market is transitioning from experiments with individual AI tools to operating entire agent systems under strict control. For business this is a signal: value now lies not only in the agent itself, but also in the observability, evaluation, and audit layer, without which scaling AI in critical processes will be too risky.