CareConnect AI is a healthcare automation system that connects incoming patient data, AI-assisted risk assessment and automated clinical alerts in one workflow. I built it with n8n as the orchestration layer and a Python/Flask service for the processing and AI logic. It started as a team project for Smart India Hackathon (SIH) 2026 with Team InnovationX.
The problem: In clinical settings, patient data often sits in different systems, and the cases that need urgent attention can be missed or noticed late. CareConnect AI automates the first-pass screening so that clinicians see the high-risk cases quickly. The system supports their decisions and does not replace them.
How the system works:
Data intake: Patient data arrives through an n8n webhook, so any form, app or hospital system can send records to the workflow with a simple API call.
Processing and validation: A Python/Flask service cleans and structures the incoming data and prepares it for assessment.
AI-assisted assessment: The service produces a risk assessment together with a confidence level for each case.
Risk- and confidence-based routing: n8n branches on both values. Low-risk, high-confidence cases follow the standard path. High-risk cases, and cases where the AI is unsure, are flagged for human review.
Automated alerts: Cases that need clinical review trigger email alerts, so the right person is notified without anyone having to monitor a dashboard.
Key design decisions:
Human in the loop: Using confidence as a routing signal means uncertain results go to a clinician, not through automatically. In healthcare this matters as much as the model itself.
Separation of concerns: n8n handles orchestration, routing and notifications, while Python/Flask handles the data and AI logic. Either side can be changed or scaled without rebuilding the other.
Webhook-first architecture: The system can plug into different front ends and data sources with little rework.
What I learned: The hardest part was making several technologies work together as one complete solution: the webhook, the Flask service, the AI layer, the routing rules and the email alerts. It taught me to design for reliability and handle edge cases like missing data, low-confidence outputs and failed notifications, on top of getting the model’s output right.
Stack: n8n, Python, Flask, AI/LLM, Webhooks, REST APIs, Email Automation
Skills tags: Workflow Automation, n8n, Python, Flask, Healthcare AI, API Integration, Webhooks
0 likes · 0 views · Nov 11, 2026
Built the automation backbone of CareConnect AI, a healthcare system that screens patient data and alerts clinicians to high-risk cases.
