What challenges did NoHarm.ai face before AI implementation?

Public hospitals in Brazil face pharmacist staffing and infrastructure constraints that lead to medication errors and preventable harm. Remote and underserved areas also lack consistent infrastructure such as electricity and internet.

What was the impact of the AI solution?

By 2025 the solution was deployed across 182 healthcare units in 17 Brazilian states, impacting more than 4 million patients. It analyzes more than 5 million prescriptions monthly and processes more than 400,000 clinical notes daily. Hospitals report up to 800% faster prescription analysis, errors fell by nearly 98% in 6 months, and one teaching hospital reduced costs by R$2 million after 2 years.

Source

ProductionEvidence: Medium65/100

NoHarm.ai prevents medication errors using Amazon Bedrock (Brazil)

Use case typeClinical decision supportUpdated Jul 15, 2026

NoHarm.ai is a healthcare technology nonprofit in Brazil that uses AWS AI to catch medication errors before they reach patients. It automates prescription review with named-entity extraction from clinical notes, contextual summarization, cross-checks against lab results, and human-in-the-loop pharmacist oversight.

Organization: NoHarm.ai
Industry: Healthcare
Location: Brazil
Published: July 2026

Reported outcomes

2,000,000 BRL

costs reducedCost savings

4 millionpatients impacted5 millionprescriptions analyzed monthly400,000 notes/dayclinical notes processed daily+800%prescription analysis speed−98%medication errors

Strategic outcomes

Other strategic outcomeExpanded safer prescription review to underserved public hospitalsRisk & complianceAdded audit trail support for regulatory complianceScale & capacityDeployed across public hospitals nationwideCost efficiencySupported a free-to-hospitals operating model

Why do we believe this?Outcome claims, sources, and evidence checks

Normalized claim

Patients impacted: 4 million increase

AWS Customer StoriesJul 15, 2026Customer storyExplicit claimMedium evidence strength

positively impacted more than 4 million patients

Normalized claim

Prescriptions analyzed monthly: 5 million increase

AWS Customer StoriesJul 15, 2026Customer storyExplicit claimMedium evidence strength

analyzes more than 5 million prescriptions monthly

Normalized claim

Clinical notes processed daily: 400,000 notes/day increase

AWS Customer StoriesJul 15, 2026Customer storyExplicit claimMedium evidence strength

processes more than 400,000 clinical notes daily

Normalized claim

Prescription analysis speed: 800% increase

AWS Customer StoriesJul 15, 2026Customer storyExplicit claimMedium evidence strength

hospitals report up to 800 percent faster prescription analysis

Normalized claim

Medication errors: 98% decrease

AWS Customer StoriesJul 15, 2026Customer storyExplicit claimMedium evidence strength

errors fell by nearly 98 percent in 6 months

Normalized claim

Prescriptions reviewed: 0.6-49% increase

AWS Customer StoriesJul 15, 2026Customer storyInferred claimMedium evidence strength

raised the share of prescriptions reviewed from 0.6 to 49 percent

Normalized claim

Costs reduced: 2,000,000 BRL decrease

AWS Customer StoriesJul 15, 2026Customer storyExplicit claimMedium evidence strength

reduced costs by R$2 million after 2 years

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: NoHarm.ai
Provider: AWS
Maturity: Production
Linked source: AWS Customer Stories

By 2025 the solution was deployed across 182 healthcare units in 17 Brazilian states, impacting more than 4 million patients

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 2 of 2

1Clinical decision support
2Healthcare workflow automation

Public hospitals in Brazil face pharmacist staffing and infrastructure constraints that lead to medication errors and preventable harm.
Remote and underserved areas also lack consistent infrastructure such as electricity and internet.

NoHarm selected AWS for flexibility, cost-effectiveness, and data residency compliance, and uses Amazon Bedrock for generative AI summarization.
The solution combines rules-based controls, named-entity recognition, risk scoring, human review, and audit trails to prioritize high-risk prescriptions and support compliance.

Technologies

Amazon Bedrock Named-entity recognition Human-in-the-loop review Risk scoring

By 2025 the solution was deployed across 182 healthcare units in 17 Brazilian states, impacting more than 4 million patients.
It analyzes more than 5 million prescriptions monthly and processes more than 400,000 clinical notes daily.
Hospitals report up to 800% faster prescription analysis, errors fell by nearly 98% in 6 months, and one teaching hospital reduced costs by R$2 million after 2 years.

Architecture

The solution combines rules-based safety controls with named-entity recognition, contextual summarization using Amazon Bedrock, cross-referencing prescriptions with laboratory results, a risk-scoring system to prioritize high-risk cases, human-in-the-loop pharmacist oversight, and an audit trail for compliance.

Sources & evidence1

Evidence: Medium65/100Evidence strength