Caretaker Dashboard

Memoral

Agentic AI for Dementia and Alzheimer’s Care

Overview

Memoral is an agentic AI companion that supports people living with dementia and Alzheimer’s at all stages. It combines conversational engagement, adaptive reminders, and intelligent task management to improve daily structure and reduce caregiver burden.

Caregivers gain clear, real-time insights into patient well-being, while patients experience consistency, connection, and compassionate interaction — all without intrusive sensors or complex hardware.

Problem Definition / User Need

Dementia and Alzheimer’s disrupt memory, communication, and daily routines. Patients lose independence, while caregivers face constant anxiety, lack of visibility, and delayed awareness of decline.

Key User Needs

Patients: Need structure, companionship, and daily reminders.
Caregivers: Need reassurance, visibility, and timely alerts.

Current Gaps:

Tools are fragmented (reminders, notes, trackers).
Few systems provide ongoing, natural interaction or actionable insights.

Solution Concept

Memoral introduces an agentic multi-agent AI system that interacts, assists, and informs.
The platform is designed around three pillars:

Engagement: Personalized conversations to support cognition and reduce loneliness.
Structure: Adaptive task and medication management.
Insight: Real-time caregiver dashboard with activity summaries and alerts.

Impact:

Improves medication adherence and daily routine consistency.
Reduces caregiver burnout through automation and transparency.
Strengthens emotional connection between patients and caregivers.

User Journey / Scenario

Morning

Memoral greets the patient and reminds them to take their medication.
If the patient delays, Memoral follows up later with a friendly nudge.

Midday

It engages the patient with conversation about familiar people or photos, promoting cognitive stimulation.

Evening

Memoral summarizes the day’s activities and sends a short report to the caregiver dashboard.
If any routines were missed, the system sends an alert.

Outcome:

The patient feels supported and connected.
The caregiver has reliable visibility without constant checking.

Mock Dialogue (Example Interaction)

Memoral: Good morning, Alice. Did you take your morning pill?
Alice: Not yet.
Memoral: No problem. Want me to remind you in ten minutes?
Alice: Yes, please.
Memoral: Great. By the way, your daughter Emma sent you a new photo. Would you like to look at it later?
Alice: That sounds lovely.

Technical Architecture

Multi-Agent AI System

Memory Agent: Provides empathetic, context-aware conversation.
Task Agent: Manages routines, reminders, and adherence tracking.
Health Agent: Observes engagement patterns and identifies changes.
Supervisor Agent: Coordinates behavior across agents and prioritizes actions.

Technology Stack

Frontend: Next.js 16 with React Server Components
AI Framework: LangGraph for agent orchestration with specialised sub-task agent
Chat Interface: CopilotKit for conversational UI
Database: SQLite for local, persistent data storage
UI Framework: TailwindCSS with Radix UI components