Beyond surveillance: why we prioritise identity over diagnosis

When we began developing CrossSense, we wanted to create a solution that would maintain and even enhance the functional capabilities of people living with dementia. Having worked in the wearable augmented reality (AR) space, we had seen its potential to elevate our abilities in the real world. Therefore, we aimed for technology that feels less like a monitor and more like a patient, understanding friend. While nearly a million people in the United Kingdom live with dementia, most tools focus on surveillance: tracking movement or alerting carers when something goes wrong. We believe that if a person living with dementia feels more confident and capable, everyone around them benefits. Technology should do much more than manage a person; it should support their ability to live as themselves and enjoy even the more difficult days.

Dementia does not cancel an individual; it disrupts specific cognitive pathways while leaving others, such as emotional experience and procedural memory, often intact and functional. Our research into multisensory processing and interventions like Cognitive Stimulation Therapy (CST) has shown that the right support, delivered at the right moment, can maintain functional independence and wellbeing for far longer than many believe. The challenge has always been delivery, as these sessions traditionally required trained facilitators. We are using state-of-the-art scientific evidence and advanced consumer technology to defy the impossible. By using an agentic AI approach, CrossSense applies these principles in real-time, appearing when a user hits a cognitive "gap", such as forgetting a step in a recipe, and then fading into the background.

CrossSense is a companion delivered through AR smartglasses. Because it sees and hears what the user does, it provides support that is Socratic rather than directive. Instead of commanding a user, the system might notice a pause while making tea and ask: "I think the kettle is ready, shall we do the next step together?". This is enabled by a multimodal backend and a knowledge graph that understands clinical evidence while learning the specific routines and objects that matter to each person individually, such as their "favourite cup".

We appreciate most types of dementia impact areas of the brain responsible for forming new memories. Nevertheless, our solution is not just a temporary fix. AR glasses anchor information in physical space, engaging multiple brain regions simultaneously to create redundancy when some neural pathways are disrupted. We use "dual coding", subtly presenting information through multiple senses to reinforce connections. This has already shown a durable learning effect in our recent product evaluation study. The CrossSense prototype helped people name objects with accuracy rising from a 46% baseline to 82% with the glasses, and up to 89% when names were shown in colour. These benefits persisted even after the glasses were removed, with participants retaining 74–78% of names an hour later. Furthermore, quality of life scores rose from 64% to 80%, with 95% of participants experiencing improvement. General cognition scores also improved after just one session. We believe our approach, once CrossSense is on the market, will slow functional decline by at least two years.

We recognise that privacy is a major concern. To address this, CrossSense's data strategy was developed within the Information Commissioner’s Office (ICO) Regulatory Sandbox. We employ a tiered consent model where the individual maintains control over what information is shared with us and their carers. Requests for specific permissions are subtly embedded in daily situations through just-in-time notices from Wispy rather than lengthy documents. This allows us to offer caregivers a nuanced, high-level overview of the day without intrusive monitoring. By handling repetitive cognitive scaffolding, CrossSense allows friends and families to focus on uniquely human connections: emotional anchoring and shared joy of being together.

The best assistive technology is that which feels invisible. Our current model uses lightweight smartglasses weighing 76g (used to be 200g+) and resembling ordinary eyewear while ever more integrating sophisticated AI capabilities. Over 250 people have contributed to shaping CrossSense to date, giving us their trust and believing in our scientific enquiry, our ethics, and the attitude of empowerment we champion. We feel a profound obligation to this community to see this project through and deliver it to the market, regardless of the obstacles we may face. Our mission is to help people lead fulfilling lives through this transformative technology, but most importantly, by rediscovering the strength within themselves.