Building a Personalised Health Assistant with AI and AWS

“Melio was a fantastic partner to work with. They were receptive to feedback, highly accommodating, and committed to ensuring a great experience.”
Tal Cook - Founder of Femmflo
Case Study
Industry
Healthcare / FemTech
Company Size
Startup
Problem Space
Women’s health literacy
Lab result interpretation
Mental wellness integration
AI health agents
Technology Used
Amazon Bedrock
AWS Lambda
Amazon S3
Amazon API Gateway
AWS CloudWatch
AWS ECS, RDS, IAM
Business Context
Women across South Africa often receive lab results filled with jargon and no follow-up support—leading to anxiety, confusion, and delayed care. Femmflo, a FemTech startup, partnered with Melio AI to launch an AI-powered chatbot that interprets medical PDFs and mental wellness data into personalised, cited responses without needing an internal tech team.
Challenge
The team needed to deliver a working, investor-ready MVP that could:
- Interpret lab result PDFs and onboarded health data
- Provide safe, cited medical insights with conversational UX
- Scale for future generative AI upgrades
- Launch within 2 months, with no in-house AI or DevOps staff
Solution
AI-enabled health assistant with AWS-native infrastructure
Melio AI delivered a chatbot MVP in 8 weeks using AWS services for agentic workflows, secure file handling, serverless compute, and scalable architecture. Key features included in the Generative AI powered Bedrock agent included:
- Context-aware answers tailored to onboarding and lab results
- Bedrock guardrails for safeguarding sensitive health-related advice for users
- Session memory and follow-up capability
- Automated fallback messages with cited sources
- OpenWebUI application hosted using AWS ECS
- Comprehensive evaluation framework to ensure the solution delivered accurate and relevant responses
The platform was designed for extensibility, with Amazon Bedrock laying the foundation for additional GenAI use cases.
Impact and Results
The MVP helped Femmflo unlock investment conversations and pilot testing. With no internal AI team, the founders were able to launch a safe, low-cost, high-impact health assistant using AWS and Melio AI’s managed stack.
⚡
Live MVP in just 8 weeks from kickoff to production
📊
Investor-ready prototype helped secure next-phase funding
💡
Personalised, cited insights from lab and wellness data
💰
Managed AWS services with no internal AI or DevOps hires needed