M_kumbusha

logo of care, warmness & inclusion
mobile responsive smart Insights with gemini

Inspiration

Vulnerable Demographic -Kenya’s elderly face chronic health risks and accessibility barriers to standard apps, while urban relatives suffer from anxiety due to a lack of visibility

Health Risks

Kenya's elderly population has surpassed 3 million, with the vast majority residing in rural areas far from specialized care.

Accessibility Barriers

Chronic illnesses like diabetes require strict adherence. Memory decline frequently compromises daily medication routines

Caregiver Anxiety

Standard apps require smartphone literacy, data, and English proficiency, alienating the rural senior population
Urban relatives lack reliable, daily visibility into the health routines of their loved ones, leading to constant worry.

What it does

Easy onboarding

both the elderly with disability like dimentia... , under a care-giver , thier routine(medication, wellbieng...) and plan of activities(doctor appoitments, massage, short exercise...) to have a profile and schedule appropiate reminders and manage by their next-of-kin

Reminder Log Analysis:

Gemini can analyze a week of logs to detect patterns (e.g., "Mrs. Jane often misses her 2 PM medication on Tuesdays").

AI Care Assistant:

A chat interface for caregivers to ask questions like "What are common side effects of Metformin for someone who is 80?"

Voice Script Generation:

Automatically generating personalized, comforting voice scripts for reminders based on the elderly person's specific needs.

Automated Reminders

Schedule-based alerts sent directly to feature phones or smartphones. No internet required for the recipient.
Localized Delivery
Voice calls in Swahili, Kikuyu, Luo, and other vernaculars, removing language barriers for rural seniors.

Dual-Alert Safety Net

If the user is unresponsive, an automated alert is instantly routed to a designated local neighbor or caregiver.

Remote Management

Intuitive web dashboard for urban relatives to schedule appointments, monitor logs, and manage subscriptions.

How we built it

webapp & USSD(without smartphone , internet) - for onmbording the care giver We built our MVP in short time focusing on speed-to-market. We used a Nodejs backend to handle API calls for our core AI feature, and Vite for a responsive frontend, google cloud console (gcp) for the generative language api. We focused heavily on user-centric design, testing with a number of early adopters to validate the workflow before scaling. The system is designed with a modular architecture, meaning we can add new features without reconfiguring the entire platform, ensuring it scales comfortably as we grow

Challenges we ran into

techinacalities Trying to get a free tire to send sms to mobile phone, since its a prototyped we used the Africa's Talking in a sandbox environment
some elderly were beyond it and the human aspect really needed to be there

Accomplishments that we're proud of

Voice-First Delivery Friendly automated calls bridge the literacy gap for rural seniors.
Vernacular Integration Swahili and regional languages (Kikuyu, Luo) for cultural trust.
Hardware Agnostic Works on basic 'kabambe' feature phones through USSD. No data plan needed.

Dual-Caregiver Loop

automated neighbor alerts create a reliable local safety net.

What we learned

you can not solve problems of everyone, but you can for some, which we exaclty solving not all Elderly people but a majority
"Thereis nothing new under the sun", Accesibility in phones already exists, social media and telecomunications exists, ours is automation and bridging the language barrier, building for local the later scale

What's next for M_kumbusha

Social Return
Directly improving the quality of life and healthcare access for rural elderly citizens.

Scalability A proven pilot model ready for expansion across other rural regions in East Africa.

Partnership Join us in creating a sustainable, tech-driven ecosystem for community-based care.
NLP (natural language processing) for maximum inclusion

More datasets integration with LLMs on local langauages, vernacular, (maasai, kikuyu, kikamba ...)