Kinabot began with a personal need—one that many of us quietly carry. As the founder, I live and work in the U.S., while my parents, both over 65, live far away. Like many families separated by distance, I often worry about their well-being, especially their cognitive health. Each time I visit once or twice a year, I wonder:

Are they remembering things clearly? Are they staying mentally sharp? Will anyone notice if something starts to change?

Through reading and research, I discovered that dementia is on the rise globally, and in most cases, it's diagnosed too late. Families often miss the earliest signs until a crisis happens. That realization struck me deeply. I needed a way to check in—not just emotionally, but cognitively—with the people I love.

And yet, in Silicon Valley, while companies are racing to build “smart” conversational robots, very few are asking what we can learn from everyday conversations. In those simple exchanges lie powerful signals—indicators of memory, orientation, language, and thought patterns.

Drawing on my background in AI research and my experience training robots at Indiana University’s AI Lab, I built Kinabot: a tool that listens not just to respond, but to understand. Kinabot assesses cognitive health through natural, daily speech—so you can stay connected to your loved one’s mental wellness, even across oceans.

KinaBot is an innovative AI tool designed to detect cognitive decline through everyday conversation.

Our founder and developer, Aoi, has worked as a data scientist in both Japan and the U.S. For years, she trained conversational AI to better understand the speech patterns of older adults. Her inspiration was deeply personal:
“I wanted to create a simple tool that even someone like my grandmother in Japan could use to check her cognitive health.”

Born from that heartfelt mission, KinaBot offers an easy, one-click way to assess cognitive function from natural speech. Combining simplicity with accuracy, KinaBot is a new kind of AI — built with care, for the aging population of Japan.

Cognitive decline affects how we speak. Research in neuroscience, linguistics, and artificial intelligence has shown that changes in cognitive health—especially early signs of conditions like dementia—manifest in patterns of spoken language. These include:

• Pauses and hesitations

• Reduced vocabulary diversity

• Repetitive or disorganized speech

• Grammatical and syntactic errors

• Changes in speech rate, pitch, and fluency

Key Scientific Findings:

1. Linguistic Biomarkers of Dementia:
   Numerous peer-reviewed studies (e.g., Fraser et al., 2016; Luz et al., 2021) have identified linguistic and acoustic features that differentiate individuals with Alzheimer's and other forms of dementia from healthy controls with over 80% accuracy using natural language samples.

2. Digital Speech Analysis is Non-Invasive & Scalable:
   Compared to traditional pen-and-paper neuropsychological tests, speech-based tools provide a non-invasive, low-cost, and repeatable alternative for early screening and ongoing monitoring.

3. AI Models Trained on Clinical Speech Data:
   State-of-the-art AI, including Natural Language Processing (NLP) and machine learning algorithms, can analyze recorded speech to detect subtle cognitive impairments with impressive accuracy. These models are trained using labeled datasets from clinical assessments and interviews.

4. Correlated with MMSE, MoCA Scores:
   Automated speech-based assessments have shown strong correlations with traditional cognitive screening tools such as the Mini-Mental State Examination (MMSE) and Montreal Cognitive Assessment (MoCA), both gold standards in cognitive evaluation.

5. Cross-Cultural and Multilingual Relevance:
   Studies (e.g., Toth et al., 2018; König et al., 2019) have demonstrated that this method works across different languages and cultural contexts, making it well-suited for Japanese-speaking older adults like those KinaBot serves.

Collecting short speech samples through natural conversation

Analyzing over 50 cognitive-linguistic features (e.g., semantic coherence, pause frequency, lexical richness)

Using proprietary machine learning models trained on both clinical and real-world data

Providing a cognitive score and visual report that reflects memory, language ability, and executive function

The name Kinabot comes from the Japanese word “Kizuna” (絆) — which means connection or bond.

This reflects our mission:
To strengthen the invisible ties between families, especially across distance, by using AI to help monitor and understand the cognitive health of loved ones.

“Kina” honors the spirit of connection, while “bot” represents the AI technology that makes it possible.
Together, Kinabot is a tool born from care, science, and the desire to stay connected—no matter where you are.

To empower families with an AI tool that mirrors cognitive health from speech—monitoring signs of dementia and supporting better care for the aging population.