This post was co-authored by Renata Salvador Grande, Gabriel Bueno, and Paulo Laurentys of Rede Mater Dei de Saúde.
The increasing adoption of multi-agent AI systems is redefining critical tasks in healthcare. In large hospital networks, where thousands of decisions directly impact cash flow, time of service, and risk of claim denials, the ability to monitor, track, and manage AI agents has become essential to operational sustainability. This is Rede Mater Dei de Saúde’s journey implementing a suite of 12 AI agents using Amazon Bedrock AgentCore, a comprehensive service that provides agent runtime, tool integration, memory management, and built-in observability for production AI agents.
About Rede Mater dei de Saude
With a 45-year history, Rede Mater Dei is one of the most respected medical institutions in Brazil, operating facilities in Belo Horizonte, Betim Contagen, Nova Lima, Salvador, Uberlandia, Goiania, Feira de Santana, and has a new project underway in São Paulo. The organization combines technology, advanced analytical intelligence, and highly complex care to deliver patient-centered outcomes and operational excellence.
Addressing structural challenges in Brazilian healthcare
According to the National Association of Private Hospitals (Anahp), claim denials in Brazil reached alarming levels in 2024, with the sector average rising from 11.89% to 15.89% and unearned revenue reaching up to R$10 billion. Like many institutions, Rede Mater Dei faced operational challenges, including:
- manual process Typically, hundreds of operational staff were on hand.
- fragmented process It was characterized by unstructured and scattered data.
- Teams with high turnover These procedures were handled by us as the repetitive nature of the work meant that staff were remote.
- complex validation The need for focused and sustained attention led to inconsistencies and rework at vulnerable stages of the process.
These weaknesses directly impacted the authentication-to-billing revenue cycle, exposing organizations to the same risks and putting pressure on increased denials across the sector. With support from A3Data and AWS, Rede Mater Dei launched a transformation program to reduce sources of rejection, accelerate analytics, and integrate observable, scalable, and high-quality operations managed through AI agents.
Suite of 12 AI agents deployed on Amazon Bedrock AgentCore Runtime
Rede Mater Dei collaborated with A3Data and the AWS Generative AI (GenAI) Innovation Center to build a program that features a complete suite of 12 AI agents designed to cover the entire hospital revenue cycle. This suite has created a “digital force” in which AI agents perceive, decide, and act in a methodical, continuous, and auditable manner, and as autonomously as possible.
The first of the 12 planned agents to be implemented are:
- contract agent: Centralize and structure complex contract rules that were previously scattered across different documents.
- parameterized agent: Automatically translate rules into a hospital’s enterprise resource planning (ERP) system to reduce human error and speed updates.
- authorization agent: Automate requests, validations, and interactions with health insurance companies.
Agent runs on Amazon Bedrock agent core runtimeprovides a secure serverless hosting environment for deploying, running, and scaling AI agents and tools.
The team organized the architecture into three complementary layers.
- DEL (Data Execution Layer): Organize data from multiple sources into a structured data lake.
- AEL (Agent Execution Layer): Coordinate and execute agents in an integrated manner.
- TCL (Trust and Compliance Layer): Apply governance, security, and compliance alignment and facilitate traceability.
To manage AI agents, Rede Mater Dei partnered with A3Data and the AWS GenAI Innovation Center. Together, they built the entire critical execution and governance layer of the agent. Amazon Bedrock AgentCorewhich became the focal point of the suite’s operations. This project is a pioneering initiative in Latin America. It tests the AgentCore evaluation on a comprehensive, large-scale AI solution for high-impact healthcare business applications.
Why choose Amazon Bedrock AgentCore?
As part of Amazon Bedrock, AgentCore is a comprehensive set of services that provides the foundation for agent use cases. It provides the modular functionality, plus tools, deployment, and observability you need to deploy, operate, and improve your AI agents at scale and securely. Components include:
Initial implementation uses AgentCore Observability and Evaluation of AgentCore.
Using AgentCore evaluation, Rede Mater Dei added the following features: Monitoring and evaluation layer of the solution. This layer supports continuous improvement. Multi-agent AI systems with measurable and controlled performance and high accuracy.
Through this service, it is possible to evaluate metrics and metrics considered to be global best practices, including accuracy, usefulness, precision, safety, objective success rate, and contextual relevance.
With this evaluation structure, Measurement and traceability For AI agents. These capabilities help maintain stability, resiliency, predictability, and regulatory compliance in healthcare environments.
Architecture using AgentCore evaluation
This solution is designed and implemented as shown below and already includes AgentCore evaluation.
Architecture AgentCore evaluation
517% ROI and structural efficiency across multiple processes
The first phase, based on AgentCore Observability and AgentCore Evaluation, enabled Rede Mater Dei to achieve tangible benefits and lay the foundation for a more secure, predictable, data-driven AI operation. In the early stages, the results were excellent both financially and clinically. 517% return on investment (ROI) in the first 4 months, Reduce authentication time by 66%and 33% reduction in surgery start time. Beyond these benefits, the AgentCore governance layer extends an agency’s ability to operate and evolve agents with control and transparency.
Governance and compliance
Structured observability provides complete traceability to key revenue cycle decisions, creating an immutable audit trail of every interaction, rule applied, and action taken by agents. This reduces regulatory risk, increases operational security, and simplifies internal and external validation, especially in sensitive processes such as contracts, authorizations, and billing.
Improved operational efficiency
With unified telemetry, teams can spend less time identifying and resolving faults, increasing Reduce incident resolution time by 50%. Teams can instantly spot anomalous agent behavior, poor performance, and inconsistencies, accelerating continuous improvement cycles and increasing reliability of workflows that directly impact rejection risk.
strategic decision making
Real-time visibility of key performance indicators (KPIs) covers automated analysis volume, expected financial impact, processing speed, estimated rejection risk, verification success rate, and metrics per insurer. These insights transformed operational data into faster and more accurate business decisions. They guide you to adjust rules, prioritize backlogs, scale teams, and surgically intervene in the workflows that most impact revenue and efficiency. Taken together, these results demonstrate that the combination of Mater Dei’s agent suite and AgentCore not only delivers immediate benefits, but also helps establish a foundation for more robust, auditable, and scalable hospital operations that can support network expansion and address Brazil’s structural challenge of claim denials.
Beyond the results achieved, this project has become a global reference and was featured in the keynote speech by Ruba Borno (VP of AWS Specialists and Partners) at AWS re:Invent 2025 in Las Vegas, demonstrating that revenue cycle transformation is not only possible, but can be measurable, fast, and generate significant benefits.
Customer testimonials
“We are working with A3Data to transform historic industry challenges with a more analytical, structured, and innovation-driven approach. Our focus is to enhance accuracy, predictability, and agility at critical revenue cycle stages, reduce volatility, and enhance operational and financial efficiency of our network. This increased consistency is naturally driven by more organizational and technically robust processes, leading to a smoother patient experience.”
– Renata Salvador Grande, Vice President Commercial and Marketing, Rede Mater Dei de Saúde
About the author
