Healthcare · AI Monitoring
Turning AI Detection
Into Caregiver Action
Real-time AI alert systems and monitoring workflows designed for senior care facilities.
Aeyesafe · October 2023
Aeyesafe is a senior care platform that helps caregivers identify and respond to potential safety concerns before they escalate.
This project focused on designing the alert experience that connected safety events to caregiver action, helping care teams identify, prioritize, and respond to incidents more effectively.
The work was completed within a six-week MVP timeline, requiring design decisions that balanced speed of delivery with the operational realities of care environments.
Opportunity
Aeyesafe's vision was to help caregivers identify and respond to potential safety concerns before they escalated.
For that vision to succeed, safety events needed to be translated into clear, actionable workflows. Caregivers needed a way to understand what happened, determine urgency, and coordinate response under time pressure.
The opportunity was to design an alert experience that could transform system-generated events into meaningful action while validating how caregivers would interact with the product in real care environments.
Solution
I designed an alert management experience that transformed system-generated events into coordinated caregiver workflows.
The solution focused on:
- • Making critical events easy to understand and prioritize
- • Establishing clear ownership and accountability
- • Supporting response workflows across multiple surfaces
- • Scaling gracefully during periods of high alert volume
Impact
- • Delivered the first end-to-end workflow connecting safety events to caregiver action
- • Reduced product risk through early validation of critical workflow assumptions
- • Created a scalable foundation for future alerting, adoption measurement, and product iteration
My Role
As the Product Designer, I:
- • Led the project from opportunity framing through research, design, prototyping, and validation
- • Defined the product strategy for how safety events would be translated into caregiver action
- • Established the ownership, prioritization, and response models that shaped the alert experience
- • Aligned user needs, operational workflows, and business objectives to deliver a pilot-ready MVP within six weeks
Given the six-week timeline, I combined domain research, competitive analysis, and caregiver feedback to quickly identify how safety events were surfaced, managed, and acted upon in existing workflows.
Three patterns emerged:
Information Was Difficult to Prioritize
Too much information made it difficult to quickly determine what required action.
Alerts Were Scattered
Critical information was distributed across multiple channels, making response workflows harder to manage.
Ownership Was Unclear
Teams could see alerts, but responsibility for responding was not always visible.
These findings showed that the challenge extended beyond alert delivery. Caregivers needed a reliable way to understand what happened, determine ownership, and prioritize response under time pressure.
My initial concept centered around a centralized alert hub that provided a single place to monitor safety events across the facility.
The hypothesis was straightforward: if caregivers had a shared source of truth, they could coordinate responses more effectively and reduce the likelihood of missed incidents.
To evaluate the concept, I conducted remote feedback sessions with caregivers and explored how alerts fit into their day-to-day workflows.
Validated
Caregivers valued having a shared source of truth for safety events.
Learned
Visibility alone did not improve response and coordination.
Direction Change
Alerts needed to support caregivers within existing workflows rather than a dedicated destination.
The feedback revealed that visibility alone was not enough to support effective response and coordination.
Across every incident, teams were forced to answer three questions before they could act:
Context
What happened?
The event, resident, and relevant context.
Prioritization
How urgent is it?
Whether the event required immediate attention or could wait.
Ownership
Who is handling it?
The current owner and response status.
Caregivers needed a way to quickly assess incoming events and determine what required attention first.
The redesigned alert experience combined contextual information, prioritization, and status visibility into a single workflow, helping teams make faster response decisions during busy shifts.
During peak periods, more than 15 alerts could be generated within a few minutes. A solution that worked well for isolated incidents quickly became overwhelming when multiple events occurred simultaneously.
The challenge was balancing urgency, visibility, and screen space without disrupting ongoing work.
Iteration A
Single Pop-up
Iteration B
Expanded List
Iteration C
Collapsible Grouping
The final approach grouped related alerts into a compact, expandable panel that kept critical information visible while preserving space for ongoing tasks.
Visibility alone does not guarantee action. Once an alert is acknowledged, teams still need a clear understanding of who is responsible and what should happen next.
The experience surfaced ownership, status, and available actions directly within the alert workflow, making accountability visible and reducing the effort required to coordinate response.
Unassigned alerts were intentionally highlighted so unresolved risk could be identified immediately.
Caregivers could assign, resolve, or dismiss alerts directly from the alert surface, keeping action close to the moment of awareness.
The final experience transformed safety events into coordinated caregiver workflows.
Caregivers could quickly understand what happened, identify urgency, determine ownership, and take action without searching across multiple systems.
By combining context, prioritization, scalable alert handling, and ownership visibility, the system helped teams move from awareness to action with less effort and greater confidence.
Pilot-Ready in Six Weeks
Moved the team from concept to a working alert system that could be tested in real care environments.
Reduced Product Risk
Validated core assumptions around caregiver workflows, alert visibility, ownership, and prioritization before launch.
Created a Foundation for Learning
Established the infrastructure needed to evaluate caregiver adoption, response behavior, and future product improvements.
This project reinforced that alerting is fundamentally a coordination problem, not a notification problem.
Detecting an event creates awareness, but value is created when caregivers can quickly understand what happened, determine urgency, and coordinate action with confidence.
The most intuitive solution was not ultimately the most effective one. Testing showed that caregivers valued a shared source of truth, but they did not want another destination that required active monitoring. The better solution was to support response within the workflows they already used.
The experience reinforced that successful alert systems are defined less by how effectively they surface information and more by how effectively they support coordinated action.
