NurseLink

Inspiration

Nurses are the operational backbone of hospitals, yet they spend only about one third of their shift in direct patient care. Frequent call light interruptions contribute significantly to task fragmentation, cognitive overload, and workflow inefficiency. On average, nurses respond to 6 to 7 patient calls per hour, many of which are non clinical in nature. At the same time, patients often wait for staff to access information about their own diagnosis, medications, or billing. Urgent and non urgent requests are funneled through the same manual system, creating unnecessary delays and stress. We asked a simple question: what if patient communication could be intelligently triaged before it ever reached a nurse? NurseLink was built to introduce an AI powered orchestration layer that separates clinical urgency from operational noise while empowering patients with direct access to their own information.

What it does

NurseLink is an AI powered hospital voice intelligence platform. Patients interact naturally with a voice assistant in their hospital room. The system: Understands intent from conversation Detects urgency using tone and sentiment analysis Classifies the request Automatically routes it to the correct department Creates structured tickets in real time Updates operational dashboards instantly There are two parallel flows:

Operational Request Flow Non clinical and clinical assistance requests are triaged and routed to: Nursing Hospitality Technical support Each request is assigned a priority score and displayed on department specific dashboards with status tracking and room level visibility.

Clinical Information Flow If a patient asks about their care plan, medications, procedures, recovery plan, providers, or billing breakdown, NurseLink securely connects to the EHR and delivers a clear, voice based response without requiring staff intervention. This reduces unnecessary interruptions while increasing patient autonomy.

How we built it

We designed NurseLink around a parallel MCP architecture.

Voice Layer A conversational AI interface captures patient speech and converts it into structured data. AI Processing Layer An MCP routing engine performs: Natural language understanding Tone and urgency detection Request classification Secure routing

Data Layer Two separated databases: Ticketing database for operational requests EHR database for patient clinical data The EHR schema is structured around: Patient profile Appointments Care plans Medications Procedures Diagnosis Recovery plan Providers Billing breakdown Dashboard Layer Nurse dashboard showing assigned patients, requests, room, department, status, and priority Hospitality dashboard for non clinical tasks This modular design ensures scalability, security, and clear separation of operational and clinical systems.

Challenges we ran into

One major challenge was designing a system that separates clinical and non clinical workflows without oversimplifying real hospital complexity.

Accurately scoring urgency based on tone required balancing sensitivity with false escalation risk.

Structuring EHR data in a way that could be securely accessed yet clearly explained in natural language also required careful schema design and response formatting.

Ensuring real time dashboard synchronization while maintaining clean database separation was another technical challenge.

Accomplishments that we're proud of

Designing a parallel MCP architecture that handles operational and clinical flows simultaneously

Implementing tone based urgency scoring to prioritize patient needs intelligently

Building a structured ticketing system with live dashboard updates Modeling a realistic EHR schema including care plans and billing hierarchies Creating a scalable system that enhances hospital efficiency without replacing core infrastructure Most importantly, we built a system that meaningfully reduces nurse interruption while preserving patient satisfaction.

What we learned

Healthcare efficiency is not just about automation. It is about intelligent orchestration. We learned that small workflow improvements at scale can have significant impact. Even reducing a portion of non clinical call volume can meaningfully increase direct patient care time. We also learned the importance of separating operational and clinical concerns architecturally while keeping the patient experience unified.

What Makes NurseLink Different

Traditional nurse-call modernization focuses on better alert delivery. NurseLink focuses on reducing unnecessary alerts altogether by structuring demand before it reaches clinical staff.

The platform acts as an operational coordination engine rather than another communication channel, enabling hospitals to:

Reduce nonclinical interruptions to nurses

Improve response routing efficiency

Capture actionable operational data that is currently lost in verbal workflows

What's next for NurseLink

Integration with live hospital EHR systems Advanced escalation logic based on longitudinal patient context Predictive workload balancing across departments Analytics dashboard for hospital administrators Multilingual voice support Our long term vision is to transform hospital communication from a reactive call light model into an intelligent, proactive orchestration system that supports both caregivers and patients at scale.