SABCHILL — DIGITAL BIOMARKER OS
Ageing is continuous.Healthcare observation should be too.
We are building the intelligence infrastructure to understand how people change as they age — through longitudinal voice and language measurements, individual baselines and clinically grounded research.
Longitudinal comparison identifies subtle temporal shifts across weeks that single episodic consultations cannot observe.
Health changes between consultations.
Healthcare often observes people at specific moments. But changes in cognition, communication and everyday function can evolve between appointments. The challenge is not simply obtaining more data. It is developing reliable ways to observe, measure and understand change over time.
Point-in-Time Clinical Snapshots
- • Episodic consultations: Vital clinical assessments occur months apart, capturing isolated moments.
- • Retrospective descriptions: Clinicians often rely on subjective recall from patients and family members regarding recent weeks.
- • Limited longitudinal measurement: Subtle day-to-day trajectories in speech timing, word retrieval, and vocal effort go unmeasured between visits.
Continuous Baseline Trajectory
- • Repeatable observations: Consent-based vernacular speech measurements captured across natural interactions.
- • Individual historical baselines: Every participant is compared against their own rolling reference distribution.
- • Structured measurement over time: Acoustic and psycholinguistic parameters tracked longitudinally under signal-quality controls.
- • Evidence linked to clinical assessments: Designed to support—never replace—clinical evaluation by physicians and researchers.
The future advantage is not a better conversation. It is what we learn across conversations.
Conversational AI is becoming increasingly accessible. The harder problem is establishing reliable relationships between repeated human communication measurements and independently verified health outcomes. We believe longitudinal, language-specific and clinically characterised data could become a meaningful foundation for the next generation of digital ageing intelligence.
Voice
Voluntary vernacular conversation
Measurement
Acoustic & linguistic signal extraction
Individual Baseline
Personalised temporal reference
Clinical Evidence
Independent clinical assessment
Intelligence
Longitudinal ageing insights
One intelligence platform. Five connected capabilities.
Human Interaction
Natural voice-based experiences in the user's preferred language.
Voice Intelligence
Analysis of speech, acoustic characteristics and language patterns.
Longitudinal Understanding
Measurement against individual historical baselines.
Clinical Evidence
Independent assessment, research validation and clinical interpretation.
Ageing Intelligence
The long-term ambition of validated intelligence supporting the understanding of ageing.
Voice Companion
Human-friendly interface and voluntary interaction
Accessible, natural-language conversations in the participant's preferred vernacular language, designed for older adults without complex application interfaces.
Technology should feel natural. Intelligence should work quietly.
The platform begins with an accessible human interface. Older adults should not need to learn complicated healthcare applications or interact with intimidating clinical technology to participate. Natural voice interaction creates the possibility of accessible, repeatable engagement.
Multilingual & Vernacular
Designed for natural Indian vernacular speech and everyday code-switching, respecting personal conversational pacing.
Distinct Consent Purposes
Ordinary companion check-ins and research-grade longitudinal health measurements operate under separate, explicit participant consent.
Every individual has a different baseline.
Age, language, dialect, education, medication, environment and natural speech variation influence how people communicate. A meaningful approach must therefore look beyond generic population averages. The platform is designed around individual longitudinal baselines, contextual measurement and clinically informed research.
Individual Baseline Band vs. Cross-Sectional Population Threshold
Methodological Note (Illustrative Only): This diagram illustrates the conceptual difference between comparing a participant against a broad cross-sectional population average versus tracking change relative to their own historical baseline. It does not depict patient outcomes or claim validated diagnostic accuracy.
Time creates context. Clinical evidence creates value.
Our long-term ambition is to build a carefully governed and scientifically characterised data infrastructure connecting vernacular speech measurements with clinically established ageing outcomes. The company's potential advantage is the quality and depth of this relationship, not the raw volume of recordings.
Purpose-Specific Consent
Participants control whether voice interactions are used solely for companion check-ins or additionally contributed to ethics-approved longitudinal research.
Architectural Separation
Identifiable operational records are isolated from pseudonymised longitudinal acoustic and linguistic feature vectors, with full withdrawal rights.
Clinical Characterisation
Longitudinal measurements gain scientific meaning only when paired with independent clinical assessments and standardized research protocols.
Build the evidence before making the claim.
We are investigating potential research applications across five structured domains in collaboration with clinical investigators:
Cognitive Ageing & Language Changes
Studying longitudinal lexical diversity, semantic density, and discourse timing in ageing cohorts.
Mild Cognitive Impairment Research
Evaluating candidate speech and hesitation markers alongside standardized neurocognitive instruments.
Parkinson's-Related Speech Changes
Investigating phonatory stability, cepstral peak prominence, and articulatory precision in motor speech research.
Mood & Well-Being Measurements
Exploring prosodic variation, pitch dynamics, and conversational engagement over time.
Functional & Multimodal Ageing Research
Studying respiratory phrase length and longitudinal functional trajectories across repeated observations.
A globally relevant intelligence challenge begins in India.
India combines a large ageing population, extraordinary linguistic diversity and an opportunity to investigate digital health measurements in populations insufficiently represented in much existing speech and health research. Our initial programme prioritises selected Indian-language environments and clinical research cohorts.
103M → 319M
India's population aged 60 and older grew from 103 million (8.6% in Census 2011) and is projected to reach 319 million (19.5% of the population) by 2050 (Bloom et al., IJE 2022).
22 / 121
22 Scheduled languages and 121 major languages documented nationally (Census of India Paper 1 of 2018), requiring native vernacular and code-switching models.
Prospective Cohort Focus
Demographic scale defines the research need—not an automatic claim of an existing dataset. Every language cohort must be built through consented prospective studies.
From voice intelligence to the science of ageing.
Vernacular Speech & Baseline Research
Vernacular speech measurements, individual baseline modelling, and longitudinal cognitive-ageing research protocols.
Clinically Validated Applications
Multi-site observational validation and clinically grounded measurement and monitoring applications.
Multimodal Ageing Signals
Integrating speech with movement, functional measurements, and other consented longitudinal health signals.
Ageing Intelligence Infrastructure
A foundational intelligence platform supporting global ageing-related research and institutional healthcare.
Understand how we age.
The next frontier is understanding how people change.
We are building a new intelligence layer at the intersection of human communication, longitudinal measurement and clinical research.