Processing...
Δ
Healthcare
AI, ML, Chatbot
Docker, NodeJS
The client, a healthcare startup, came with an ambitious but straightforward business idea—create a high-load chatbot performing as a personal medical assistant based on behavioral profiling and AI personalization. The aim was to increase patient engagement and optimize workflow management in medical organizations. The idea came out of statistics that nearly 50% of patients don’t take medication prescribed by doctors—this significantly influences overall healthcare outcomes.
The client had preliminary agreements with huge medical organizations that provided a ready-to-use database of over 420,000 patients to adopt the soon-to-be chatbot. The startup’s leadership had chosen NIX United as a power-packed and reputable vendor capable of coping with challenging tasks.
The main challenge was to create a high-load and customizable chatbot with multiple service scenarios like identifying patients, learning their demographic characteristics and habits, producing personalized communication, and as a result, synchronizing them under one tech umbrella.
Managing up to 8,000,000 users with 1,200,000 parallel active chat sessions
Processing 250 requests per second with 0.5-second average response time
Following medical security regulations and standards like FHIR and HIPAA
NIX offered to accelerate the entire process with a team-as-a-service (TaaS) approach. The client was satisfied with the provided tech expertise during the first milestone and team involvement; for this reason, leadership departed from micromanagement. TaaS comprises a well-running and skilled team to drive the project development autonomously so the client can save time and focus on strategic business aspects, not draining much of his own resources.
A certain client’s control over the work and total transparency is saved due to direct dealing with a project manager. After implementing the TaaS approach, product development progress skyrocketed.
To make the chatbot smart, our team reinforced it with AI, but to make it a full-fledged personal healthcare assistant, we made some additional improvements. The functionality included detecting available healthcare services relevant to the patient, scheduling appointments, making automatic arrangements at hospitals, sending notifications to take medication or pick it up from the pharmacy, and more. To deploy this complex project, a series of steps were taken.
The team created a complex architecture to cover the logic of the whole chatbot system.
The system is represented by three interfaces applying to the workflow execution subsystem and internal services as the core of the entire system. The performance of MVP was tested on 50,000 users. In one week, 420,000 real users were added to the platform.
The product had to be dynamic and agile. The NIX team created 15 microservices, each running processes with its own logic and messaging with its own scenario. Horizontal scaling of microservices was enabled by using Docker containers in combination with Kubernetes. Now each component is independently deployable, highly maintainable, and testable.
Azure autoscaling was used to automatically decrease the number of active servers running behind the platform when user activity is low and increasing server capacity when there are spikes in activity. Our solution made the system more flexible and saved a lot of money for the client.
Customization was a hard requirement as each healthcare organization has its own software, specific tools of management, and different numbers of patients. NIX created a set of comprehensive workflows enabling quick integration with any type of medical organization. Each workflow was written as pseudocode describing various internal processes and logic to change the dynamic behavior of the chatbot without additional code or downtime.
To secure data, the team followed stringent industry-standard regulations like FHIR and medical information security standards like HIPAA. All personal data and medical profiles were encrypted, transmitted, and stored separately to eliminate any possibilities of data leakage.
With our help the client successfully launched an AI-powered chatbot for the 420,000 customer base. The initial tech requirements for 8 million users are achievable due to microservice architecture and autoscaling capabilities. Our tech solutions saved time and money for the leadership of the startup. We continue collaboration with the client to develop a mobile app for further product scaling.
The AI-based chatbot has already provided a measurable impact. The startup’s leadership fulfilled all the preliminary agreements with the healthcare organizations. TaaS enabled the client to focus on the strategic aspects and business deals. As a result, more medical institutions conducted agreements to integrate the highly customized chatbot into their admin portals. The chatbot is a highly demanded solution in the healthcare industry:
This will all significantly improve the overall health system in the world.
Project Manager, Business Analyst, QA Lead, QA Automation Engineer, Manual QA Engineer, Software Architect, Tech Lead, 5 JS Developers, 2 DevOps Engine
NodeJS/TypeScript, ReactJS, NestJS, Docker, Kubernetes, Terraform, Gitea, Harbor, Vault, Jenkins, Ansible, SonarQube, Anthos ConfigSync, PostgreSQL, Redis, Azure Service Bus / RabbitMQ, HTTP Rest, GRPC, WebSocket, Azure Blob Storage, Firebase, Twilio
Schedule Meeting
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.