LastingMind helps people preserve their stories, values and relationships, so the family who come after them can revisit those stories, and one day talk with them.

The first version asked for a vast amount of deeply personal information with no sense of where to start, how the activities connected or where it all led. People finished onboarding not knowing what the app was for, and then they left. I traced it to four problems.

I studied similar products and read the science of remembering: memory, nostalgia, motivation and reminiscence. Then I turned what I found into four principles that every screen had to meet.

I reorganised the app into three phases that move from familiar facts toward deeper reflection, plus My Study, which is always open for everyday journaling. New phases unlock as you go, so there’s always a clear next step, and never the whole journey at once.

The home screen holds it together in three parts.

A growing tree stands for a legacy taking shape. Stars acknowledge every contribution and show its depth: each topic can earn up to three. Reaching a star threshold unlocks the next part of the app and grows the tree, so steady, low-pressure progress becomes something you can see.

Because the subject is family, memory and mortality, I left out streaks and competition in favour of quiet, encouraging signs of growth.

The original flow pitched features, then dropped people straight into the hardest question there is: “How would you like to be remembered?” I rewrote the first few minutes as a story: a seed that becomes a tree, a few gentle personal questions, and a clear picture of who it’s all for.

Voice makes it easier to contribute, but people still need to know what to talk about, how much is left, and whether the AI understood them. I explored three models, and each showed a different strength.

The final direction, the Guided Conversation, asks one focused question at a time, then shows the names, relationships and details the AI picked out, for the person to check before anything is saved. Instead of asking people to trust that it understood, it shows them.

Different memories need different kinds of interaction. I designed three voice-first formats, with typing always available, so the journey never turns into one long conversation. Quick answers also give the AI context it can come back to in deeper reflections.

I designed LastingMind’s brand and built its design system: the mark, the palette, the type, the illustrations and the components every screen is made from.

Older people warmed to a rich, storybook look over a clean, minimal one, so the brand is warm: a tree, a cream page and a serif with character. One palette, two typefaces and one set of parts keep it simple to use.

I built the prototype with AI, in Windsurf and Claude Code, over about four weeks, working from written specs for the app’s structure, its unlock rules and its reward flows. It has about 30 screens in React, TypeScript, Vite, Tailwind CSS and Framer Motion, runs on mock data and scripts its conversations.

Engineering built the production iOS app from it. The design kept moving as I built; here’s how the first spec compares with what shipped.

LastingMind 2.0 launched on the App Store in August 2026, its first public release. In the first month it reached 300+ users, and weekly engagement doubled.

In four weeks, an open-ended concept became a mobile product with a clear journey, several ways to contribute and a progress system people can feel.