Turning everyday conversation into a cognitive performance score

Cognitive Healthtech arrived with two years of research and a proprietary framework for reading cognitive performance through natural communication patterns. It worked, but it was manual, dependent on interpretation, and impossible to run at scale.

The framework had to become software: automated and repeatable, scalable enough to carry pilots, partnerships and clinical work without a rebuild each time, and able to work with the conversations people actually have rather than a test taken in a lab. The founding bet was that the richest signal is already there in everyday speech, every call, meeting and moment of reflection.

We built the platform end to end. Phase 1 turned the framework into a working pipeline: consented capture of audio from Microsoft Teams, Google Meet and direct upload; automated transcription and extraction of dozens of conversational signals; and a daily, person-centred cognitive score broken into five interpretable dimensions, communication flow, cognitive focus, affective expression, social reciprocity and memory. Every score carries a confidence rating, and the dashboard (WCAG 2.1 AA) is deliberately supportive and non-diagnostic.

Privacy and ethics were designed in from the first sprint, not retro-fitted: pseudonymisation, encryption in transit and at rest, UK-hosted infrastructure, consent management and full audit logging, built against UK GDPR throughout. The MVP was piloted with a real cohort, producing both the first real-world performance data and the footage that trained Phase 2.

Phase 2 took it from prototype to product. The indicators were re-trained on roughly 100 hours of the platform's own Phase 1 footage; score articulation now shows a user why a score moved; and behavioural video intelligence lets them jump from a dashboard insight straight to the second in a session that drove it. It ships as native iOS and Android apps, with Stripe subscriptions and single-use vouchers for trials and research cohorts, and a configurable clinical pathway that adapts to regulatory guidance without rewriting the product.

Cognitive Healthtech was awarded one of a handful of places on the UKRI STFC Health Mission Accelerator for dementia, a competitive, government-backed programme aligned to the NHS Fit for the Future strategy, with the 2026 cohort running from January to March.

Behind that sits a working AI scoring engine across five dimensions with confidence ratings, models custom-trained on the platform's own data, an MVP validated with real users, and a mobile-ready product heading for public release. BearTech continues as the build partner as the platform moves toward larger datasets, a dedicated clinician dashboard, and AI-driven recommendations.

Read the full case study, the extended version with more detail on each phase.