Back An AI-Powered Wealth Planning Platform Case Study
image 301
Problem Statement Retirement planning is genuinely hard to design for. People have accounts scattered across different institutions, transfers take forever and feel opaque, and there's always this low-level anxiety about making a tax mistake or running out of money. The goal was to design a hub that brings everything together in one place and actually makes people feel safe and in control, not like they're filing their taxes.
My role I am the only designer and researcher for end-to-end UX/UI design process
Period 5 Days
How I Used AI in This Project Used AI tools: Polymet, UX Pilot, and ChatGPT

I used AI tools early in the process to generate a quick structural baseline and surface common UI patterns fast. That compression of early exploration was genuinely useful. What normally takes days of initial sketching happened in hours, which meant I could spend more time on the thing that actually matters: whether the experience holds up emotionally for someone who's nervous about their retirement savings.

I want to be clear about what that means in practice though. The AI gave me a starting point. It did not give me a design.
Yandy Hung
Research and Design Process
Persona Linda Thompson is 66, a recently retired nurse from Mississauga. Her income comes from CPP, OAS, a RRIF, and a modest TFSA, all scattered across different institutions. She's not financially clueless, she just never had everything in one place. Tax rules around RRIF withdrawals confuse her and she's afraid of making a costly mistake. She wants one clear picture of her finances and the confidence to manage it without constant stress.
image 291
Journey map Mapping Linda's journey across five stages showed where anxiety concentrates. The pain points clustered early, around disconnected accounts and unclear processes. The opportunities that came out of this weren't about more features. They were about reducing confusion at the moments that mattered most.
image 292
These two artifacts shaped every major decision. When I could see exactly where Linda's frustration peaked, the design responses became obvious: a consolidated hub, plain-language guidance, and status visibility throughout the transfer flow.
What the AI Got Wrong This is the part I found most useful about the experiment. Within minutes of generating the initial UI, I could see exactly where AI breaks down on financial products:
螢幕截圖 2025-09-15 下午8.55.17 1
Lack of hierarchy Flat, overwhelming layout. Everything had equal visual weight, like a survey form. There was no sense of what mattered most or what the user should focus on first.
Component
Jargon-heavy language User Jargon-heavy language without explanations.
Component
Cold, transactional tone Missing empathy. No warmth, no reassurance, nothing that acknowledges this is a stressful thing to do with your money.
image 298
Missing emotional design & reassurance no reassurance during waiting periods or supportive messaging.
image 299
What I Fixed and Why
These were not polish decisions. Each one came from thinking about what it actually feels like to be a 60-year-old moving your retirement savings to a new platform for the first time.
  • I restructured the hierarchy because financial decisions need breathing room. When everything competes for attention, nothing gets trusted.

  • I rewrote everything in plain language because jargon is a trust killer at the moment it matters most. If someone has to Google a term mid-flow, you've already lost them.

  • I added empathetic microcopy because a waiting screen during a fund transfer is not a neutral moment. It's anxiety. Good design acknowledges that.

  • I redesigned the answer choices to be contextual because asking someone to self-identify their caregiver status using government definitions, without any guidance, is a design failure, not a user failure.
Prototype of human enhanced version
Transfer Flow - Start with clicking the “Transfer” on left menu bar, then “Start Conversion” button on the banner
Start with account details: enter the institution, account type, and number; a stepper shows progress and inline checks catch errors. Upload the required documents with clear statuses and simple error messages, then answer a short, plain-language KYC with tooltips and easy choices. Review everything on one page, edit in place, and submit to a confirmation screen with a real-time status tracker and timelines.
Investment Builder - Start with clicking the “Retire Investment Builder” on left menu bar
Enter a few basics—age, expected income, and savings—with optional advanced settings in collapsible panels. Take a quick, jargon-free risk check using simple multiple-choice cards. See projected balances and tax-aware withdrawal tips in clear charts, then review, adjust, or accept the plan with a digital acknowledgment (disclaimers shown for transparency).
What I Took Away AI made me faster at the beginning. My judgment is what made it worth shipping.
  • The most valuable thing this project showed me is that AI tools are genuinely useful for compressing early exploration, but they have a real blind spot: they don't understand what it feels like to be the user. They don't know that handing over your retirement savings to a new platform is emotionally loaded. They don't know that "your money is safe" is not obvious, it needs to be said. That's the designer's job. That's always going to be the designer's job.
An AI-Powered Wealth Planning Platform Case Study
AI Tools Case Study
Yandy Hung
Back
image 300
My role I am the only designer and researcher for end-to-end UX/UI design process
Period 5 Days
Problem Statement Retirement planning is genuinely hard to design for. People have accounts scattered across different institutions, transfers take forever and feel opaque, and there's always this low-level anxiety about making a tax mistake or running out of money. The goal was to design a hub that brings everything together in one place and actually makes people feel safe and in control, not like they're filing their taxes.
How I Used AI in This Project Used AI tools: Polymet, UX Pilot, and ChatGPT

I used AI tools early in the process to generate a quick structural baseline and surface common UI patterns fast. That compression of early exploration was genuinely useful. What normally takes days of initial sketching happened in hours, which meant I could spend more time on the thing that actually matters: whether the experience holds up emotionally for someone who's nervous about their retirement savings.

I want to be clear about what that means in practice though. The AI gave me a starting point. It did not give me a design.
Research and Design Process
Persona
Linda Thompson is 66, a recently retired nurse from Mississauga. Her income comes from CPP, OAS, a RRIF, and a modest TFSA, all scattered across different institutions. She's not financially clueless, she just never had everything in one place. Tax rules around RRIF withdrawals confuse her and she's afraid of making a costly mistake. She wants one clear picture of her finances and the confidence to manage it without constant stress.
image 291
Journey map
Mapping Linda's journey across five stages showed where anxiety concentrates. The pain points clustered early, around disconnected accounts and unclear processes. The opportunities that came out of this weren't about more features. They were about reducing confusion at the moments that mattered most.
image 292
These two artifacts shaped every major decision. When I could see exactly where Linda's frustration peaked, the design responses became obvious: a consolidated hub, plain-language guidance, and status visibility throughout the transfer flow.
What the AI Got Wrong This is the part I found most useful about the experiment. Within minutes of generating the initial UI, I could see exactly where AI breaks down on financial products:
螢幕截圖 2025-09-15 下午8.55.17 1
Lack of hierarchy Flat, overwhelming layout. Everything had equal visual weight, like a survey form. There was no sense of what mattered most or what the user should focus on first.
Component
Component
Jargon-heavy language User Jargon-heavy language without explanations.
Cold, transactional tone Missing empathy. No warmth, no reassurance, nothing that acknowledges this is a stressful thing to do with your money.
image 298
image 299
Missing emotional design & reassurance No reassurance during waiting periods or supportive messaging.
What I Fixed and Why These were not polish decisions. Each one came from thinking about what it actually feels like to be a 60-year-old moving your retirement savings to a new platform for the first time.
  • I restructured the hierarchy because financial decisions need breathing room. When everything competes for attention, nothing gets trusted.

  • I rewrote everything in plain language because jargon is a trust killer at the moment it matters most. If someone has to Google a term mid-flow, you've already lost them.

  • I added empathetic microcopy because a waiting screen during a fund transfer is not a neutral moment. It's anxiety. Good design acknowledges that.

  • I redesigned the answer choices to be contextual because asking someone to self-identify their caregiver status using government definitions, without any guidance, is a design failure, not a user failure.
Prototype of human enhanced version
Transfer Flow - Start with clicking the “Transfer” on left menu bar, then “Start Conversion” button on the banner
Start with account details: enter the institution, account type, and number; a stepper shows progress and inline checks catch errors. Upload the required documents with clear statuses and simple error messages, then answer a short, plain-language KYC with tooltips and easy choices. Review everything on one page, edit in place, and submit to a confirmation screen with a real-time status tracker and timelines.
Investment Builder - Start with clicking the “Retire Investment Builder” on left menu bar
Enter a few basics—age, expected income, and savings—with optional advanced settings in collapsible panels. Take a quick, jargon-free risk check using simple multiple-choice cards. See projected balances and tax-aware withdrawal tips in clear charts, then review, adjust, or accept the plan with a digital acknowledgment (disclaimers shown for transparency).
What I Took Away AI made me faster at the beginning. My judgment is what made it worth shipping.
The most valuable thing this project showed me is that AI tools are genuinely useful for compressing early exploration, but they have a real blind spot: they don't understand what it feels like to be the user. They don't know that handing over your retirement savings to a new platform is emotionally loaded. They don't know that "your money is safe" is not obvious, it needs to be said. That's the designer's job. That's always going to be the designer's job.