AI PRODUCT STRATEGY · RESPONSIBLE AI · PROTOTYPE

Amicus Vera

A calm AI companion and family-support experience designed to help older adults feel connected, supported and in control.

THE OPPORTUNITY

Support should feel like companionship—not surveillance.

Older adults can need help with connection, routines and moments of uncertainty, while families need reassurance that respects independence. Existing experiences can feel fragmented, clinical or intrusive.

Product thesis Dignity before automation: use AI to make support feel more human, while keeping people in control.

DECISION BOUNDARIES

AI can assist. Rules and people retain control.

AI

Assist and propose

Draft short conversational responses and propose a memory item. The user can ignore, correct or reject the output.

RULES

Control permissions

Consent, sharing, deletion, reminders and SOS routing remain deterministic product logic—not model decisions.

HUMAN

Own high-stakes decisions

Qualified people and emergency services handle medical interpretation, emergency decisions and serious safety incidents.

CORE JOURNEYS

Designed around real moments of need.

01

Daily connection

A familiar, low-pressure check-in helps an older adult share how they are feeling and move through reminders without a clinical tone.

02

Concerning moment

The proposed AI experience responds calmly, avoids diagnosis and offers a clear path to a trusted human or emergency support when needed.

03

Family reassurance

A consent-led family view provides useful signals without turning support into constant surveillance.

AI ROLE

Useful context, careful response and clear escalation.

The product concept specifies how AI could understand conversational context, respond in a warm and accessible way, support continuity and recognise when a human should take over.

Important boundary: this is a product and interaction specification demonstrated through a clickable prototype—not a claim of a production AI system or live model integration.

01 · ContextUnderstand the user's words and current journey02 · ResponseUse calm, clear and non-diagnostic language03 · EscalationOffer trusted human support when risk appears04 · ContinuityRemember only what the user has agreed to share

MY CONTRIBUTION

From ambiguous need to inspectable product direction.

01Problem framing and product thesis02Personas, needs and end-to-end journeys03Feature prioritisation and requirements04AI behaviour and safety boundaries05Prototype direction and QA06Success measures and staged roadmap

WHAT EXISTS TODAY

Evidence, stated precisely.

Demonstrated

  • A public, clickable product prototype
  • Defined users, journeys and product principles
  • Specified AI behaviours and responsible-use boundaries
  • Version 0.2 product documentation, evaluation gates and roadmap

Not claimed

  • Live database or production AI integration
  • Clinical validation or medical-device status
  • Institutional deployment or measured outcomes
  • Production security or regulatory readiness

NEXT VALIDATION

Test comprehension, comfort and safety before adding complexity.

5–8older-adult usability sessions3–5family-member interviews0acceptable critical failures in safety scenarios

These are proposed research targets, not completed studies or reported outcomes.

Portfolio truth statement: Amicus Vera demonstrates product strategy, intended journeys and a clickable interaction concept. It is not a live AI model, database, clinical product, production security implementation or institutional deployment.

EXPLORE THE WORK

See the product thinking in action.