Rayyan Shaikh← All projects
Case study · AI companionship

MeloMate

An AI companion designed to ease loneliness in an ageing population, offering emotional support while protecting privacy, and fostering connection without replacing human relationships.

Role
UX Lead
Year
2025
Team
6 designers
Context
Uppsala University · MSc HCI
MeloMate · an AI companion for an ageing population
MeloMate · an AI companion for an ageing population
§ 01The problem

Loneliness is rising among older adults, and AI companions promise to help. But today's companions are tuned so closely to the user that they become a mirror, which risks dependency and erodes real connection.

01
Privacy & trust

Older users are hesitant to share personal thoughts with AI because of data security and who might access their information.

02
Authentic, not addictive

Companionship must feel real without fostering dependency or replacing human relationships.

03
Personal but independent

Balancing personalisation with an AI that has its own character, rather than one that simply echoes the user.

We set out to design a user-friendly companion tailored to older adults, with controlled, customisable interactions and ethical data management at its core, grounded in a literature review and a survey of the people it was for.

§ 02My role · UX Lead
  1. 01
    Literature review & survey

    I led the user research: a qualitative survey in English and Swedish (Google Forms & email) covering attitudes toward AI companions, social connections, current technology use and demographics. We designed it around GDPR's principles: informed consent, anonymity, no intrusive questions and the right to withdraw at any time.

  2. 02
    Analysis

    Most respondents anticipated social isolation as they age. Half were comfortable sharing personal thoughts with AI, while a third hesitated over privacy. Text messaging was the preferred way to interact, 75% wanted the AI to adapt its personality over time, and practical features like daily reminders were favourites.

  3. 03
    Storyboard, lo-fi & MoSCoW

    We sketched storyboards and paper prototypes of the core journey (log in, create an account, meet your companion), then ran a MoSCoW analysis to agree what the app must, should, could and won't have (for example: no commercial advertising).

  4. 04
    Hi-fi prototype & evaluation

    The hi-fi prototype was evaluated and refined: we fixed an inconsistent colour scheme, added AI voice features, removed the mandatory address field for privacy, expanded the privacy and ethics sections and added a dedicated terms page.

§ 03From the process
Survey results · comfort sharing with AI and preferred interaction
Survey results · comfort sharing with AI and preferred interaction
Expectations: adjustable personality and practical features
Expectations: adjustable personality and practical features
Storyboard, lo-fi prototype and MoSCoW analysis
Storyboard, lo-fi prototype and MoSCoW analysis
Hi-fi user flow and evaluation changes
Hi-fi user flow and evaluation changes
Final design · sign-up and privacy-first onboarding
Final design · sign-up and privacy-first onboarding
Creating a MeloMate in five guided steps
Creating a MeloMate in five guided steps
§ 04Outcome

A companion that grows with you, without becoming you.

  • Privacy-first onboarding: “your privacy, our responsibility” explained before any data is asked for.
  • A five-step MeloMate maker with common ground and some choice, but not unlimited customisation, so the AI's personality can develop over time.
  • The companion lives inside existing messaging apps, blending with real human conversations rather than standing apart as another screen.
§ 05Reflection

“MeloMate is where my interest in human–AI interaction got personal. We were so preoccupied with whether AI could be a companion that the real design question became whether and how it should.”