Improving financial well-being through AI assistants

Product Design
A mobile mockup of GoodWealth on a minimalistic background.

The Problem

Financial Literacy is at an all time low among adults which directly impacts financial wellbeing. After initial research, we found that Gen Z (aged 18-25) averaged 43% on a financial literacy test, the lowest financial literacy scores among generations, but with the highest desire to increase their financial literacy and overall financial well-being. With around 20% of the population being Gen Z, increasing financial literacy through education can help banks acquire and retain customers. I formulated a How Might We question based on this to frame the process forward.

The Benefit

Increasing financial literacy through education can help acquire and retain Gen Z customers.

I formulated a How Might We question based on this to frame the process forward. To tackle this question, I had to focus on three main areas:

  • Financial Education

  • Managing Money

  • Go-to Sources

How Might We Question

“A young adult's (age 18-25) problem is that they lack sufficient financial knowledge to confidently make decisions. How might we help them gain knowledge to make smart financial decisions and increase their financial well-being?”

The Solution

To tackle these three areas, I conducted a survey gathering insights as to how people currently learn about finances, how they manage their money, and what sources of financial information they use. So findings include:

  • 40% learned through their parents.

  • 80% are dissatisfied or somewhat dissatisfied with their current level of financial literacy.

  • 80% agree that they want to have a better understanding overall of finances.

  • 40% want to learn through an advisor or personal course.

I then conducted a user interview to get an understanding how users actually behave and feel towards managing their own finances and mapped out the findings on an empathy map. With the HMW question in mind, I rapidly ideated solutions to the problem.

An empathy map made an user interview.
Empathy Map
Final Solution

Personal financial advisor assistant. Helps with budgeting, personal and investment advice.

A table synthesizing quotes from user interviews to design features.
Synthesizing design requirements from user interview.

Through information from the empathy map, I identified problem areas which can be translated into design solutions for each main area of focus.

  • Financial Education → Chatbots & Financial testing

  • Managing Money → Financial Assistants

  • Go-to Sources → Financial Advice

I mapped out a flowchart to illustrate the flow of the system. This also helped map out where any technical solutions needed to be implemented.

Flowchart of the system.
A mobile mockup of GoodWealth's chatbot feature.

Chatbots & Financial Testing

I designed a system for financial education that:

  • Allows users to ask questions related to a specific financial area.

  • Allows users to ask any question related to finances.

  • Bases tests off of knowledge rather than memory by generating test questions. 

A mobile mockup of GoodWealth's financial assistants.

Financial Assistants

I also created three financial tools to help users track their money.

  • Budget Assistant generates a unique budget plan based on the user's financial situation.

  • Expense Assistant tracks expenses and generates a report to adjust spending habits.

  • Investment Assistant manages the users portfolio and gives predictions. 

A mobile mockup of GoodWealth's financial advice feature.

Financial Advice

Lastly, getting financial advice can be accessed through chatbots or assistants, which all cross reference each other to give holistic and specific financial advice.

Technical Requirements

To get a technical understanding of implementing this system, I researched further possible solutions for the features. We largely have four technical requirements that need to be addressed:

  • Conversational AI for Chatbot

  • Generate data visualization and graphs

  • Generate questions for financial testing

  • Access to third party banking platforms

For our chatbot and test question generating, we need to use an LLM like ChatGPT or Cohere. Using ChatGPT or Cohere gives us the best flexibility for scalability and security over any financial information. We can easily generate data visualizations using a third party service like Chartify or ChartAI.

Technical map for GoodWealth features.
Technical Map for different app features.

Limitations & Considerations


  • Contextual Advice - AI can be trained to give advice on finances, but also needs to learn about user end goals, attitudes, behaviours, and their financial situation to give useful financial advice.

  • Prompting the System - Prompt engineering is relatively new. Since it's the user generating the prompt, there’s also the assumption that the user already knows the problem, as opposed to human advisors who can ask questions to diagnose the problem.


  • Gaining Trust - Trust is at the heart of every financial decision. The product needs to be designed in a way that users can trust the AI with their financial livelihood.

  • Financial Nuance - Large financial decisions are situational and nuanced. There’s rarely any black and white.

  • Learning through Generative AI -Understanding how people learn and the impact of learning through artificial intelligence is something that needs to be further researched.

  • Premature for Gen Z - People in Gen Z are still young and might not be in financial situations where they need to think about these financial topics.


This project taught me how to design and technically implement an AI based product. Next steps for this project would be to conduct usability testing and iterate based on user feedback.

A mockup of multiple mobile screens designs for GoodWealth.
GoodWealth Mobile Screens