Internship
Product Design
Fintech
Capybara.AI
PROBLEM
The vast amount of stock information on the internet and the volatile stock market make it difficult for people to make the proper stock selections by getting complete information on time.
GOAL
As the design lead of this company, I was expected to design an AI-powered information software from scratch. Besides that, I’m also going to collaborate with PM and developers to decide future functions, conduct user testing, and launch the product.
TEAM
1 leader, 6 designers, 10 front-end & back-end developers
DURATION
2023.Jun ~ 2024. Jan (7 months)
Things I did
40+
Reusable design components
70+
User Interviews
20+
User Testings
Impact
20%
Lift of the agile development process effectiveness
30%
Information processing time shortened
Sponsored by Microsoft for startups and NYU. The first version will be launched in late March 2024.
Process behind Capybara.AI
Market Analyze | User Research | Function Hierarchy
Competitive Analyze | Design System
User Testing | Iterations
How to stand out from other stock investment products on the market?
Market analysis
Will people interest in a product like Capybara.AI?
Initial user research
People who have 0~3 years and 9+ years of experience in finance industry are interested in a product that combines AI and finance.
Stock investment products on the market only push countless pieces of information to the user. They haven’t consider how the user will process those complex information.
Also, there is a lack of analytics and evaluation features tailored to the user. This make it hard for users to start investing confidently.
Target User Groups
How to get closer to our future users, know their real feelings and needs?
User Research Plan
To better specify user’s pain points, define user groups, determine future product directions and functionalities, I made a research plan to conduct user research and send the survey to 104 people.
User Interviews & Key Insights
Design Challenge
How to design a software that is rich in information and features, enabling users to access the most comprehensive information in the shortest time, alleviating tension during investment, and distinguishing it from other stock investment applications?
Ideation
I generated these four ideas based on the three main directions discussed with the leader. Following conversations with both the leader and the developers, I confirmed that all these functions can be implemented.
Product Record Document
After conducting card sorting and mapping out the information structure with the leader, I created the product record document containing the following information:
Product background: target users, market analysis, purpose
Functions: reasons for having each function, purpose, advanced/basic features, entries, interaction flow
Iterations: record of every version of the app, reasons for edits
Due to the complexity and size of this application, I aim to ensure a clear and organized structure.
Version 1 of Capybara.AI includes:
All functions
Complete user flows
Design details that are not yet perfected
How to find out usability problems in this application with numerous functions?
User Testing Plan
This round of user testing aims to refine the user flow for critical functions, discover potential user needs, and identify user interface problems.
Each participant is assigned several tasks that cover key functions of the product. I observe and time them while they complete the tasks. Afterward, I ask several follow-up questions based on their behavior.
Focus Group
5 retail investigators
3 financial professionals
Retail Investigator’s Feedback
Things to Improve
Save stock function need to be more complete.
There seems to be an excess of information on each page.
The use of bright colors is everywhere, which makes it a bit distracting.
Things to Keep
Risk score and sentiment score are valuable tools for users, helping to alleviate concerns and fears during trading.
The design of the pie chart is really straightforward, displaying and categorizing information clearly for the users.
The explanations of professional terms are helpful for retail investigators.
Financial Profession’s Feedback
Things to Improve
Consider expand your AI analyze tool to other fields, such as social media. Make your product a tool to help people know more about a company, not only stocks.
If the sentiment score and risk score do not have a 100% accurate prediction of stock movement or whether users will incur losses, it is advisable not to add biased colors, as it may mislead people.
Things to Keep
Breaking news notifications are helpful in keeping users informed even while they are busy working.
The design of weekly investment poster is very clear and well structured. Help me save time to get the information I need.
UI & Interaction
How design plays an important role in helping the product achieve its goals and distinguishes it from other competitors.
Design System
High-fidelity Pages
Design Delivery
After finishing the prototypes in Figma, I wrote a detailed design delivery document that includes UX flows and interaction details for front-end developers. To address design problems and ensure everything stays on track, I also meet with developers once a week
Next Steps
Short-term: After development
Bug bash the software demo.
Online & offline user testings.
Short-term: After launch
Collect online user feedback weekly.
Initiate more professional accessibility testing.
Long-term: APP Version 2
Add at least 5 new functions to the app and re-organize information architecture.