Turning market noise into answers for retail investors

Collaborator

Yilin Wang

3 Product Managers

8 AI Engineers

Timeline

2023 - 2025

Scope

Brand → AI product design

Key design metrics

What the design moved

User retention

40% / 32% / 24%

Day 1, Day 7, and Day 30 retention.

Follow-up rate

20.8% → 35.3%

After adding AI data visualization skills to the conversational AI feature.

View Ask Capy

Daily active users

10K

DAU within one month of launching the MVP.

Design system

40+

Reusable components built for the product.

View design system

Problem

Retail investors have more financial news than ever, but no easy way to connect it to their own holdings.

To understand a single market move, they jump between news sites, brokerage apps, and social media, then piece together the story themselves.

Solution

Catch up on what happened within a minute

Open the app to see how your portfolio moved, and the news behind it.

Understand what the news means for you

Ask Capy turns the numbers into clear answers, so you get the story at a glance.

Research

What 20+ investor interviews told us about who we’re designing for

Working with the founder and 3 engineers, I used internal AI tools to synthesize transcripts from 20+ retail investor interviews, and the outcome revealed who we are designing for and their respective pain points.

Investors found news comes from multiple sources all the time. Because their holdings are scattered and various, it’s harder to find the related news.

Design Opportunity

synthesize all the information online, surface what matters to investors.

User Painpoint 1

Retail investors often struggle to get a full picture of market changes in a short time.

News card wireframe — headline, sentiment, time frame, summary

To tackle this problem, we decided to pull news from verified financial sources and then structure the summaries by companies.

Design Challenge

How might we surface the information that matters most to the retail investors?

Exploration

News grouped by company

Display news summaries by company, sorted by each stock’s weight in the investor’s portfolio

Usability Testing

Are we showing users what actually matters?

SourceFeedbackDesign next step
Retail investor AmyRetail investors also follow stocks they don’t own or watch.Include broader market news and high-volatility stocks.
Retail investor LukeUsers read headlines first and check price performance second.Prioritize the news headlines, make price data secondary.
AI engineerMajor news can affect several companies at once.Visualize the relationship between a single news and multiple tickers.
AI engineerNews about the same stock vary in importance and sentiment.Rank by individual news priority, surface sentiment.

Solution

News feed cards, one per headline, each with its tickers and sentiment

Display by each headline, sort based on the impact to the investors’ portfolios and the market.

News detail page with an Also impacted row of META, NVDA and TSLA

In each news individual page, investors can see the other stocks impacted by the single news.

User Painpoint 2

Retail investors know something changed, but not what it means for their portfolio and what to do next.

Ask Capy draws on the portfolio, viewed stocks, top news and live market data to suggest questions and structure answers

To tackle this problem, we designed a conversational AI assistant “Ask Capy”.

Design Challenge

How and where should Ask Capy appear to stay discoverable while reducing cognitive load?

Exploration

A dedicated tab in the bottom navigation bar.

Usability Testing

Users’ questions come up while reading news, tracking stock prices, and viewing their portfolio performance, not when they first open the app.

Solution

Add an Ask Capy section with question prompts that tie directly to the user’s portfolio, their stocks, and the day’s top news.

Ask Capy prompts on the Portfolio, stock and Explore screens

AI & Design

However, the AI output is very lengthy and dense. How to make AI answers easier to scan and understand?

Therefore, I built an AI skill that turns key data comparisons into visual charts.

AI output pipeline: design-time rules feed a run-time loop that fills, validates, repairs and renders answer cards

After applying the AI skill, the output contains data visualizations that make it easier to understand.

Design System

Brand & Visual Identity

Building the brand, visual language, and design system from scratch to give a new product a credible, trustworthy voice.

Design system — light theme
Design system — dark theme

Key learnings

What I took away

  1. 01

    The system’s logic isn’t the user’s logic.

    I learned to question designs that feel natural to the system but not to the user.

  2. 02

    The best moment to offer help is the moment the question arises.