2.5M
World Bank
Thai people over 80 projected to need assistance after a more than six-fold increase.
Research Visualization
Thai dementia care is a family system under pressure.
The research points to one narrow MVP: Thai voice check-in that turns stressful home-care moments into a structured care log, caregiver burden signal, safe next step, and family handoff.
Decision Build, but narrow
User Adult child caregiver
Boundary Wellness support, not diagnosis
01Thai voice check-in
02Structured incident log
03Caregiver burden signal
04Safe next step
05Family handoff summary
Why The Problem Is Real
Four numbers make the case quickly.
These are the first statistics to show judges or collaborators before discussing product features.
600K → 2M
BMJ Open
Estimated dementia population path in Thailand: about 600K in 2015, potentially 2M by 2050.
86%
BMJ Open
Rural Thailand dementia caregivers in one study were household informal caregivers.
80%
ADI 2024
General public share in ADI 2024 report who think dementia is a normal part of ageing.
Caregiver Burden Surface
The product should track burden as a first-class signal.
The strongest pain point is not missing information. It is repeated stressful care episodes that become sleep loss, family conflict, isolation, and burnout risk.
Sleep Collapse
Night waking, exit-seeking, and repeated checking turn one incident into a full-day wellness problem.
Stress And Guilt
Thai family duty can make asking for help feel like blame or failure.
Work And Time
Caregivers may also work outside the home, making handoff clarity and task sharing important.
Money Pressure
Institutional long-term care coverage is limited, so families often manage at home longer.
Sibling Conflict
Unstructured updates make it easy for relatives to misunderstand risk and effort.
Hidden Health Risk
The caregiver's own physical and mental health becomes part of the care system.
Thai Family Care Reality
Care happens inside a family workflow, not only a clinic workflow.
The MVP should respect family duty while making burden visible enough to coordinate help.
Parent At Home
Dementia symptoms appear in everyday routines, often outside clinical visits.
Adult Child Watches
The adult child notices patterns, safety changes, sleep disruption, and ADL friction.
Stress Builds
Duty, exhaustion, and uncertainty accumulate before anyone formally asks for help.
Siblings Need Facts
Relatives need concise evidence, not emotional arguments, to share responsibility.
Professionals Need History
Clinicians or hired caregivers benefit from a clear timeline if the family shares it.
Social Listening Pattern
Real care moments are messy, emotional, and spoken.
The demo should start from natural Thai speech, then visibly transform it into structured support.
Demo Input Moment
เมื่อคืนแม่ตื่นหลายรอบ จะออกจากบ้าน บอกว่าจะกลับบ้าน หนูแทบไม่ได้นอน
This input is not a clean form submission. It is the exact kind of tired, compressed story where voice and structured outputs become useful.
Wandering / exit-seekingHigh
Repeated questionsHigh
Sleep disruptionHigh
ADL resistanceMed
Caregiver exhaustionHigh
Current Response Gap
Thailand has useful support, but not a continuous home-care co-pilot.
The opportunity is not replacing hospitals, NGOs, or paid caregivers. It is connecting the stressful moments between them.
Hospitals
Trusted Thai education and clinical care, but not always available during nighttime family stress.
Education Pages
Helpful dementia guidance, but usually static and not personalized to the current incident.
Care Services
Home-care providers and marketplaces help families buy labor, often in urban/private settings.
Family Chat
Fast and familiar, but emotionally loaded and poor at building reliable care history.
AI Co-Pilot
Thai voice, caregiver wellness, incident structure, safe guidance, and family handoff in one workflow.
Competitive White Space
The gap is the combination, not one isolated feature.
Competitors cluster around services, education, coordination, patient safety, and generic AI wellness. The white space is Thai-first caregiver wellness plus dementia-specific structure.
Caregiver Wellness Centered
Patient / Service Centered
Low Thai Localization
High Thai Localization
OnCare
ElderThai
Thai hospital pages
Alzheimer's Association
CogniCare
Medisafe
Wysa / Woebot
Lotsa Helping Hands
Thai Caregiver Wellness AI
Strong local fit, but service-heavy and not a self-service AI support layer.
Trusted and useful, but mostly passive and not built around live family handoff.
Conversationally strong, but not Thai dementia-care-specific or family-system aware.
Thai-first voice plus structured outputs for caregiver burden, safety, and family coordination.
OpenAI API Spine
The demo should make the transformation visible.
Hero API feature: Realtime voice plus Structured Outputs. Optional later: file or vision input for non-diagnostic context.
Realtime Voice
Caregiver speaks naturally in Thai during a stressful moment instead of filling a form.
Follow-Up
AI asks only the missing safety and burden questions: safe now, left home, sleep hours.
Structured Outputs
The story becomes parseable fields: incident type, risk, sleep, burden score, family ask.
Safe Guidance
Non-diagnostic next steps and escalation boundaries appear in Thai-first wording.
Family Handoff
AI drafts a sibling message that separates facts, risk, burden, and a concrete help request.
3-5 Minute Demo
Show one emotionally clear workflow from stress to handoff.
Seeded demo data is acceptable for reliability. The OpenAI-powered transformation should be the visible center.
Voice Check-In
Daughter describes mother waking repeatedly, trying to leave home, and caregiver sleeping only three hours.
Incident Log
System creates a wandering / exit-seeking record with high safety risk and high caregiver burden.
Safe Next Steps
Confirm safety, avoid arguing about reality, prepare contact info, ask family to cover rest time.
Sibling Handoff
Thai message asks for a specific shift or task while showing facts and caregiver burden.
Weekly Trend
Repeated sleep and wandering incidents trigger a family meeting and professional discussion suggestion.
Safety Boundary
Strong positioning keeps the MVP credible.
This must be framed as caregiver support, documentation, family coordination, education, and escalation guidance.
The AI Can Do
Support the caregiver in reflection, logging, education, and communication.
Voice check-in
Incident logging
Burden tracking
Family handoff
Education
Escalation guidance
The AI Cannot Do
Replace clinicians, emergency services, or medication decisions.
Diagnosis
Treatment
Medication changes
Emergency replacement
Clinician replacement
Sedative advice
Final Build Decision
Build the MVP. Keep it narrow, Thai, caregiver-first, and OpenAI-visible.
The research supports a focused hackathon prototype, not a broad eldercare platform.
Recommended MVP
Thai voice check-in -> structured incident log -> caregiver burden signal -> safe next step -> family handoff summary.
Thai-first
Dementia-specific
Caregiver wellness
Family coordination
Non-diagnostic
Realtime + Structured Outputs
Do Not Build In V1
A broad eldercare marketplace, a clinical dementia management platform, a medication adviser, or a generic mental-health chatbot.
Narrowing protects demo clarity
Source Trail
Primary sources used in the visualization.
World BankThailand aging and long-term care pressure.
BMJ OpenRural Thailand dementia care and informal caregiver reliance.
JMAT / ThaiScienceThai dementia caregiver burden and needs.
WHOGlobal dementia public-health response and carer support.
ADI 2024Global dementia attitudes, stigma, and misunderstanding.
OpenAI RealtimeVoice check-in as the hero interaction.
OpenAI Structured OutputsReliable care logs, risk fields, and family handoff schema.
OpenAI VisionOptional non-diagnostic context from care cards or labels.
OpenAI File InputsOptional hospital instructions or routine documents.