GlucAssist - Transforming diabetes management with generative AI agents on AWS
Location
Peru
Industry
HealthcareSolution
Generative AITechnology(ies)
Amazon Bedrock, AWS Scheduler, Amazon RDS, Amazon S3, Amazon API Gateway, AWS Amplify
Company
GlucAssist, a Peruvian healthcare startup, is revolutionizing diabetes control and monitoring through a generative AI-based virtual agent that operates via WhatsApp.
Together with Amber, they implemented a serverless architecture on AWS that leverages the power of Amazon Bedrock to create natural interactions with patients, intelligent reminders, and automated reports.
The result: a significant improvement in treatment adherence, monitoring frequency, and quality of life for diabetes patients.
Challenge
Diabetes affects more than 6% of the Peruvian population and requires constant monitoring to prevent serious complications.
GlucAssist detected that patients faced difficulties such as forgetting medication, lack of organized records, and limited communication with doctors.
Our proposal
Amber developed an intelligent conversational agent with AWS technology that operates through WhatsApp Business, allowing patients to interact in natural language, record data, and receive personalized automatic reminders.
The system uses generative AI on Amazon Bedrock to understand user context and respond empathetically and accurately.
The solution runs on a 100% serverless architecture, ensuring scalability, security, and optimized costs under a pay-per-use model.
Key technology components:
- Amazon Bedrock: Natural language processing with generative AI.
- AWS Scheduler: Automatic management of personalized reminders.
- Amazon RDS: Secure storage of medical records.
- Amazon S3: Storage of reports and glucometer images.
- Amazon API Gateway: Integration with WhatsApp Business API.
- AWS Amplify: Web portal for configuration and administration.
Implementation process
The project was developed over 4 months using an agile methodology with 2-week sprints, prioritizing constant feedback from patients and doctors.
Phase 1: Discovery and design
Workshops with the GlucAssist team, patient interviews, and interaction prototyping.
Phase 2: MVP Development
Creation of the virtual agent and connection with AWS services.
Phase 3: Testing and refinement
Validation with real users and conversational language adjustment.
Phase 4: Launch
Final deployment, full automation, and continuous support.
patients
- 89% medication treatment adherence (vs. 47% previously).
- 73% increase in glucose monitoring frequency.
- 62% more regular physical activity recorded.
- 40% fewer diabetes-related emergency visits.
doctors
- Access to more complete and reliable data.
- Automated reports for better decision-making.
- Constant and fluid communication with patients.
- Optimization of consultation and follow-up time.
GlucAssist
- Market launch in just 4 months.
- Scalable and flexible architecture for future growth.
- Optimized cost model with real pay-per-use.
- Technological base adaptable to other chronic conditions.
Testimonials
María Lopez Type 2 diabetes patient
"Since I started using GlucAssist, I feel accompanied in my treatment. The WhatsApp reminders are like having a personal assistant that understands me and adapts to my needs."
Dr. Carlos Ramirez Endocrinologist
"The ability to review system-generated reports allows me to make better medical decisions for my diabetes patients. I can see clear trends and adjust treatments with greater precision."
Hugo Siu Co-founder of GlucAssist
"With Amber and their AI expertise, we managed to create a simple, approachable, and effective solution for the people who need it most. Today GlucAssist is not just an app, it's a life tool."
Other success stories
We drive business efficiency with artificial intelligence agents that automate processes and deliver real results.
