Back Aidan Liu
App Developer · UX Designer

Elderly AI App

An iOS companion app designed for users 65 and older. Built, shipped, and published to the App Store in 10 weeks.

Jun - Aug 2025 App Store GitHub
1 Live iOS app
4 Core screens
65+ Target users
10 Weeks to ship

Designed for clarity, built for confidence

9:41 🤖 ElderlyAI Welcome to ElderlyAI 👋 📖 Learn Basics 🤖 Ask AI Helper 📰 What's New
Home
9:41 Ask AI Helper Hello! I'm here to help. Ask me anything about using your iPhone. How do I install an app? Easy! Open the App Store, search for the app, then tap "Get" to download it. ⭐ Save ▶ Play response Type a question...
AI Assistant
9:41 Settings FONT SIZE 16 20 24 28 SPEECH SPEED Slow Normal Fast YOUR PROGRESS Guides read 3 Guides completed 1 Favorites saved 5 ⭐
Settings

Four features built for people who find most apps overwhelming

Learn Basics
Step-by-step guides on everyday tech: writing emails, video calls, using the web. Written in plain language with no jargon, no assumed knowledge.
Ask AI Helper
GPT-3.5 powered assistant with text-to-speech output. Users speak or type questions and hear answers read aloud at an adjustable pace.
What's New
A curated, plain-language feed of AI news. No algorithm, no clutter. Relevant developments summarized clearly for everyday readers.
Adaptive Settings
Font size (16-28px), speech speed (Slow/Normal/Fast), and full progress tracking. Every preference persists so users never have to reconfigure.

Zero to App Store in 10 weeks

Jun 2025
Ideation
Market research, user interviews, feature scoping with the team
Jul 2025
Build
React Native in Expo, GPT-3.5 API, text-to-speech integration
Aug 2025
Beta
65+ beta testers, 2 App Store rejections, privacy policy revisions
Aug 2025
Live
Published to App Store, ASO optimization, first organic downloads

Overview

AI built for the people who need it most

The idea came from noticing how much of the existing AI landscape is built for tech-savvy users. Tools like ChatGPT assume you know what a prompt is. We wanted to build something for older adults who might benefit the most from AI but find most interfaces overwhelming.

We focused relentlessly on one question: can a 70-year-old who has never used a smartphone understand this immediately?

"My 74-year-old mother can actually use this. Finally an app that doesn't assume everyone is 25."

App Store Review
65+ average age of target user

Design Decisions

Every pixel was decided by asking one question

Can a 70-year-old who has never used a smartphone understand this immediately? That meant large buttons, 20px body text as the floor, high-contrast white cards, and exactly three primary actions on the home screen.

We shipped without a hamburger menu, a buried settings icon, or any gesture-based navigation. All primary actions are reachable in one tap from home.

Design Rules
👆
Touch targets 44pt min
Aa
Body text floor 20px
☀️
Contrast ratio WCAG AA
🏠
Primary actions 3 max
🚫
Gestures required None

Design is not just visual. It is understanding your user deeply enough to anticipate their confusion before it happens.

Aidan Liu — what I learned from this project

Five things that were harder than expected

01
Getting the conversational tone right
Most LLM outputs feel clinical or overly verbose. We spent weeks prompt-engineering responses to feel warm, patient, and appropriately paced for older adults.
02
App Store review rejections
Two rejections before we got the privacy disclosures and data handling documentation right. Shipping taught me how much work happens outside the code.
03
Testing with real 65+ users
Most design assumptions fail immediately when you watch a real elderly user try your app. We ran four rounds of in-person beta testing to unlearn our own biases.
04
Font size and layout flexibility
Supporting four font sizes (16/20/24/28px) without breaking every layout required systematic design thinking, not ad-hoc fixes.
05
Text-to-speech speed calibration
The default system TTS speed is too fast. We built a three-speed system and let users discover the right speed themselves rather than picking a default.

What I Learned

First time taking something from zero to the App Store

The biggest lesson: design is anticipating confusion before it happens. The second: the gap between "it works" and "it ships" is enormous, and most of that gap has nothing to do with code.

I also learned how to gather qualitative feedback from non-technical users, navigate App Store policy, and think seriously about user acquisition and retention.

01
Shipping teaches you things building doesn't. Review, user responses, and edge cases revealed more than any dev environment could.
02
Design is a user research problem. Every visual decision was downstream of observing real users fail at something obvious to us.
03
Business matters even in a side project. ASO, user acquisition, App Store screenshots determine whether anyone finds what you built.
04
Moving fast is hard. Decisions happen quickly but require constant alignment. Communication is the most important feature.
Available Now

Try ElderlyAI on iPhone