Resilience Early Screening for Independent Living in the Community
RESILIENCE is a community-based frailty screening and intelligence system built for the NHS. It identifies older adults at risk before crisis, gives them something of genuine personal value from the process, and equips clinical teams with the intelligence they need to respond effectively.
There is a population of older adults in the UK who are too frail to be fully safe at home but not yet eligible for formal NHS or social care support. They fall between the thresholds. They are one fall, one infection, or one carer breakdown away from an acute hospital admission, often avoidable.
The NHS frailty system is reactive. People appear in it when they arrive at A&E or when crisis has already happened. Proactive community identification of frailty risk, before the system has to respond, barely exists at scale anywhere in the country.
For every older adult who loses their independence to an avoidable crisis, there is a family under strain, a carer invisible to health services, and a preventive opportunity that passed without anyone seeing it. RESILIENCE is designed to see it.
RESILIENCE asks older adults about their everyday life in a conversational, voice-first interface powered by Claude Sonnet. They speak freely, in their own words, at their own pace. The AI listens: not just to what is said, but to how it is said. It detects hesitation, minimisation, and the specific fears that prevent older adults from disclosing vulnerability: the fear of being a burden, and the fear of what honest disclosure might lead to.
From a single conversation, RESILIENCE generates a personalised Wellness Guide for the person, a structured clinical referral for the frailty team, and, with consent, a population intelligence signal for NHS commissioners. Three distinct outputs. One dignified, human conversation. Nothing shared without explicit permission.
A voice-first community frailty screen that interprets free spoken responses using Claude Sonnet. Twelve conversational questions across six clinical domains. Adaptive language for four completing-party types. Safety architecture built throughout.
Open tool →An independent observational questionnaire for a nominated carer, linked by consent only. The carer's scores never reach the person; the person's scores never reach the carer. The comparison happens at the clinical layer, where it generates the most value.
Open tool → SimulationA structured referral document for the frailty team, centred on a Frailty Profile Radar, a six-spoke diagram that communicates the person's risk profile at a glance. Domain detail, minimisation notes, referral priorities, and a CGA preparation brief.
This tool shows a simulation of how clinical referral outputs will be presented. It is not yet connected to live screening data.
View simulation → SimulationWhen both a person's screen and a carer's report are available, two profiles overlay on the radar. Agreement confirms clinical confidence. Divergence is a clinical signal, almost always indicating that the person is understating genuine difficulty.
This tool shows a simulation of how dual person and carer outputs will be presented side by side. It is not yet connected to live screening data.
View simulation → SimulationA population-level frailty intelligence interface for NHS commissioning teams. Geographic ward-level risk mapping, domain analysis, referral demand forecasting, and an AI-synthesised commissioning brief from the full anonymised dataset.
This tool shows a simulation of how population intelligence outputs will be presented to commissioning teams. It is not yet connected to live screening data.
View simulation →The synthesis layer. Person self-report, carer observation, and CGA clinical findings overlaid on a single radar, three independent perspectives on the same person. In this simulation, where all three confirm significant need, the Triple Tap is the highest confidence signal in the system.
View simulation →RESILIENCE gives before it asks. Every person receives a personalised Wellness Guide, regardless of what they choose to share with clinical services. The guide belongs to them.
You are the author of your own life. Nothing happens without the person's permission. Every consent layer is granular, revocable, and explained in plain language.
Everyone wants the same thing. The older person, the carer, and the clinical team all want the same outcome: safe, well, and at home, on the person's own terms. RESILIENCE closes the information gap between them.
The AI is always under human control. RESILIENCE uses Claude Sonnet to interpret spoken responses and generate outputs, but the AI operates within a tightly defined set of instructions written by the development team. It does not act freely: it follows a clear clinical framework, asks only the questions it is given, and interprets responses only within the domains it is designed to address. This is a working prototype, and the instruction set will be refined in direct partnership with geriatricians, frailty nurses, and clinical researchers. The system learns through that collaboration, not independently.
Privacy is built in, not bolted on. RESILIENCE was designed from the outset to be GDPR-compliant by default. No personal data is stored, transmitted, or shared unless the person explicitly chooses to share it. The three-layer consent model gives each individual granular control: a person can receive their Wellness Guide and share nothing further. Data used for population intelligence is anonymised before it leaves the device. There is no profiling, no third-party data sharing, and no commercial use of personal information. Privacy is not a feature of RESILIENCE. It is the foundation.
British Geriatrics Society self-screening frailty identification instrument, embedded invisibly across the twelve screening questions and scored in real time.
Fatigue, Resistance, Ambulation, Illness, and Loss of weight domains mapped across Physical Function, Nutrition, and Medical Burden question categories.
British Geriatrics Society framework informing question categories, clinical domain structure, and referral routing throughout the tool suite.
Finnish Geriatric Intervention Study: multidomain intervention evidence base underpinning the personalised Wellness Guide structure and recommendations.
NHS RightCare Frailty Pathway informing clinical triage logic, referral routing, and priority weighting within the Clinical Referral tool.
Peer-reviewed evidence covering oral frailty indicators and sleep disruption as supplementary clinical domains within the screening architecture.
A 2025 peer-reviewed study in Geriatric Nursing (Zhang et al., DOI: 10.1016/j.gerinurse.2024.10.025) tested nine machine learning models for predicting frailty in older adults. The best-performing model identified pain level, depression, and functional ability as the three strongest predictors, outperforming objective physical measures. Notably conducted in a chronic pain population, the finding that subjective, self-reported experience carries greater predictive weight than clinical history directly supports RESILIENCE's conversational, voice-first approach to eliciting what standardised forms cannot.
RESILIENCE is a working prototype, developed in response to a real clinical need and grounded in validated evidence frameworks. The clinical domain structure and question architecture have been informed by informal conversations with geriatric specialists and frailty practitioners. The next stage is formal co-development: working directly with those at the cutting edge of delivering support and interventions to this population, to test, challenge, and strengthen every aspect of the system.
RESILIENCE is for you if you want to stay living well at home and would like to understand what support might help, on your own terms, at your own pace.
RESILIENCE is for frailty nurses, geriatricians, Neighbourhood Health Service coordinators, and GP practices who need to identify at-risk patients in the community before crisis, not after.
RESILIENCE is for ICB commissioners and NHS system leaders who need population-level frailty intelligence to inform commissioning decisions, identify service gaps, and target preventive investment.
RESILIENCE is not a clinical assessment and it is not a form. It is a conversation about your everyday life. It will ask you about things like how you are sleeping, whether you feel confident getting around, and what matters most to you about staying at home. At the end, it gives you a personalised Wellness Guide that belongs to you. Nothing is passed on to anyone unless you choose to share it.
Maybe things feel a little harder than they used to. Maybe you are managing fine but would like to understand what support is available before you need it. RESILIENCE asks you about your life in your own words, at your own pace, and gives you something genuinely useful in return. No waiting room. No referral needed. Just a conversation.
Try the conversation →Adult children often notice things before anyone else does. A parent who seems a little more tired, a little less steady, a little more reluctant to talk about how they are getting on. If you have noticed something and are not sure what to do next, RESILIENCE can help you understand the picture more clearly, and give your family member something valuable in the process.
Learn how it works →Social prescribing link workers, community pharmacists, voluntary sector workers, and faith community leaders are often the first to sense that someone is struggling. RESILIENCE gives you something to offer in that moment: a dignified, voice-first conversation that produces a Wellness Guide the person can keep, and, if they choose, opens a pathway to clinical support.
Talk to us about your work →"The guide is yours. We give it to you regardless of what you choose to do next. Nothing leaves this conversation without your permission."
RESILIENCE is developed by Assistiv Systems. If you are a clinician, researcher, or specialist working in this field and would like to explore the prototype or offer feedback, we would welcome the conversation.
simon@assistiv.co