The Reality
I started by mapping the experience of a typical small business owner dealing with finances. Not the happy path. The real one.They open their accounting tool once a month, when something goes wrong. They see a number they don’t recognize. They click around. They give up.
What I Kept Hearing (And Observing)
Users didn't avoid their finances because they didn't care, they avoided them because every interaction felt like a test they hadn't studied for.
"I open my accounting app maybe once a month. Usually when something's already gone wrong." - Freelance photographer, San Francisco
"I know the numbers are there. I just can't tell what they're actually telling me."
- Small business owner, 6-person team
"Every tool I've tried assumes I already know what I'm looking for. I don't. That's the whole problem." - Independent consultant
The problem isn’t missing features. It’s that the entire information architecture assumes the user already knows what they’re looking for.
My Role
UX Designer
Duration
1 month
Responsibilities
User Research, UX Planning,
UI Design and Prototyping
Link
Selected Works
Flo
A concept product that makes business finances feel less like a chore. Flo is an AI-powered finance tool where you just ask questions and your numbers talk back.
Overview
Most small business owners aren’t bad at money.
They’re bad at software.
A freelancer running $80k in revenue shouldn’t need an accountant just to understand why their cash felt tight in January. But that’s the reality. Xero, QuickBooks, FreshBooks - they give you data. They don’t give you answers.
What if your finances just… talked to
you?
Solution
That’s the question Flo was built around. Not a smarter dashboard. A conversation.
Target Users
First-time Finance Manager
Age 21-30
Behaviour Markers
• Recently started earning or saving seriously • Tries to track expenses but gets inconsistent • Feels overwhelmed by financial terms
Pain Points
Doesn’t know how to interpret or act on financial data
Freelancer
Age 22-35
Behaviour Markers
• Manages income from multiple sources • Checks balance often but avoids deep analysis • Uses basic tools like notes, UPI apps, or spreadsheets
Pain Points
Irregular income makes it hard to track and plan
Small Business Owner
Age 28-45
Behaviour Markers
• Juggles expenses, revenue, and operations • Looks at numbers only when needed • Relies on accountants but lacks day-to-day visibility
Pain Points
No clear, real-time understanding of business health
Competitive Insights
What Makes Flo Different
There's a gap between:
Data (QuickBooks/Xero)
Visibility (Mercury)
Understanding (missing layer)
First Wrong Direction
I built the dashboard first. That was the mistake.
My initial thinking was: people need a better overview screen. Cleaner KPIs. Better charts. A “health score” that gives a single number at a glance.
I got fairly deep into it. Four KPI cards, a cashflow chart, an expense breakdown by category. It looked good. It felt competent. It was exactly what everyone else had already built.
Why It Failed
It was still passive - the user has to interpret, compare, and conclude.
The “health score” idea collapsed quickly: what counts as ‘healthy’ varies wildly by business type.
I was solving for aesthetics, not for the underlying friction.
A cleaner dashboard is still a dashboard.
The Turning Point
The reframe came from a single question: what if the first thing you see isn’t a graph, but a response to something you asked?
Suddenly the whole structure changed. The chat interface isn’t a feature it’s the entry point. The data panel on the right exists to expand on what Flo said, not the other way around.
You ask. Flo answers. Then you dig in. That sequence completely changed how I thought about the layout.
Key Insights
People don't need more data. They need a trusted voice.
The core design principle for Flo - it should feel like a conversation with a calm, knowledgeable advisor not a query interface or a report generator.
What This Meant In Practice
Flo's response comes first, in plain language not tables, not chart.
The data card in chat is a compressed preview, not the full story.
The 'Flo's Insight' strip adds interpretation, not just display.
Suggested follow-ups guide the user, they shouldn't have to think of what to ask next.
The visual identity needed to feel calm and trustworthy, not cold and technical
Product Thinking
Product UI - The Experience
The main design challenge: how do conversation and data coexist on one screen without one cannibalizing the other? The answer was spatial separation with intentional linking. The chat panel stays narrow and conversational. The data panel is wide and analytical. They're connected by the preview card in chat which is a compressed version of what's expanded on the right.
Trade Offs: What I Chose Not To Build
No onboarding flow - the brief was one product screen, not a journey. I focused depth over breadth.
No empty states - everything is shown in a populated, mid-use state to communicate value faster
No mobile views - the product UI is desktop-only by design. The financial analysis use case maps to focused, desktop contexts.
Tensions I Navigated
Conversational AI vs. navigable data - resolved with the two-panel approach: chat for dialogue, panel for exploration.
Warm editorial identity vs. technical AI product - resolved with the later iteration: Syne font (geometric, architectural) replaces Playfair, teal replaces sage, but warmth is preserved in spacing and tone.
Showing Flo’s intelligence without over-promising - the ‘Insight’ strip is labeled clearly; Flo explains its reasoning, not just its conclusions
Design Execution
I chose a conversational-first layout over a dashboard because the mental model shifts from 'reading data' to 'getting an answer' that sequence changes everything about information hierarchy.
DM Mono for all financial values creates a visual language that separates data from prose, the user always knows when they're looking at a number vs. an explanation."
The three-column split was deliberate, chat stays narrow so it never feels like a spreadsheet. The data panel only exists to deepen what Flo already said.
Expected User Outcomes
This is a concept product, so there are no live metrics. But the design communicates specific, testable hypotheses:
Expected User Outcomes
<30s
Time to Understand Expenses
+40%
Navigation Efficiency
+35%
Insight Awareness
-30%
Cognitive Load
Reflection
What I'd Do Differently
The thing I’m most uncertain about: the chat history as the primary navigation model. Having five recent conversations in the sidebar feels right for a power user who asks Flo things daily. But for someone who opens the app once a month? The sidebar could feel empty in a way that undermines the product.
I’d want to test two variants: one where the sidebar defaults to recent chats (current design), and one where it defaults to smart suggestions, Flo proactively surfacing things worth asking about before you even type anything.
What I Learned
The hardest part of designing a conversational product is resisting the urge to also make it a dashboard. They want different things from you.
The data panel was the right call, but its relationship to the chat panel took several iterations to feel natural rather than forced.
My Role
UX Designer
Duration
1 month
Responsibilities
User Research, UX Planning,
UI Design and Prototyping
Link
Overview
Most small business owners aren’t bad at money.
They’re bad at software.
A freelancer running $80k in revenue shouldn’t need an accountant just to understand why their cash felt tight in January. But that’s the reality. Xero, QuickBooks, FreshBooks - they give you data. They don’t give you answers.
What if your finances just… talked to you?
Solution
That’s the question Flo was built around. Not a smarter dashboard. A conversation.
The Reality
I started by mapping the experience of a typical small business owner dealing with finances. Not the happy path. The real one.They open their accounting tool once a month, when something goes wrong. They see a number they don’t recognize. They click around. They give up.
What I Kept Hearing (And Observing)
Users didn't avoid their finances because they didn't care, they avoided them because every interaction felt like a test they hadn't studied for.
"I open my accounting app maybe once a month. Usually when something's already gone wrong." - Freelance photographer, San Francisco
"I know the numbers are there. I just can't tell what they're actually telling me."
- Small business owner
"Every tool I've tried assumes I already know what I'm looking for. I don't. That's the whole problem." - Independent consultant
The problem isn’t missing features. It’s that the entire information architecture assumes the user already knows what they’re looking for.
Target Users
First-time Finance Manager
Age 21-30
Behaviour Markers
• Recently started earning or saving seriously • Tries to track expenses but gets inconsistent • Feels overwhelmed by financial terms
Pain Points
Doesn’t know how to interpret or act on financial data
Freelancer/ Solo Creator
Age 22-35
Behaviour Markers
• Manages income from multiple sources • Checks balance often but avoids deep analysis • Uses basic tools like notes, UPI apps, or spreadsheets
Pain Points
Irregular income makes it hard to track and plan
Small Business Owner
Age 28-45
Behaviour Markers
• Juggles expenses, revenue, and operations • Looks at numbers only when needed • Relies on accountants but lacks day-to-day visibility
Pain Points
No clear, real-time understanding of business health
Competitive Insights
What Makes Flo Different
There's a gap between:
Data (QuickBooks/Xero)
Visibility (Mercury)
Understanding (missing layer)
First Wrong Direction
I built the dashboard first. That was the mistake.
My initial thinking was: people need a better overview screen. Cleaner KPIs. Better charts. A “health score” that gives a single number at a glance.
I got fairly deep into it. Four KPI cards, a cashflow chart, an expense breakdown by category. It looked good. It felt competent. It was exactly what everyone else had already built.
Why It Failed
It was still passive - the user has to interpret, compare, and conclude
The “health score” idea collapsed quickly: what counts as ‘healthy’ varies wildly by business type
I was solving for aesthetics, not for the underlying friction
A cleaner dashboard is still a dashboard
The Turning Point
The reframe came from a single question: what if the first thing you see isn’t a graph, but a response to something you asked?
Suddenly the whole structure changed. The chat interface isn’t a feature it’s the entry point. The data panel on the right exists to expand on what Flo said, not the other way around.
You ask. Flo answers. Then you dig in. That sequence completely changed how I thought about the layout.
Key Insights
People don't need more data. They need a trusted voice.
The core design principle for Flo - it should feel like a conversation with a calm, knowledgeable advisor not a query interface or a report generator.
What This Meant In Practice
Flo's response comes first, in plain language not tables, not charts.
The data card in chat is a compressed preview, not the full story.
The 'Flo's Insight' strip adds interpretation, not just display.
Suggested follow-ups guide the user, they shouldn't have to think of what to ask next.
The visual identity needed to feel calm and trustworthy, not cold and technical.
Product Thinking
Product UI - The Experience
The main design challenge: how do conversation and data coexist on one screen without one cannibalizing the other? The answer was spatial separation with intentional linking. The chat panel stays narrow and conversational. The data panel is wide and analytical. They're connected by the preview card in chat which is a compressed version of what's expanded on the right.
Trade Offs: What I Chose Not To Build
No onboarding flow - the brief was one product screen, not a journey. I focused depth over breadth.
No empty states - everything is shown in a populated, mid-use state to communicate value faster
No mobile views - the product UI is desktop-only by design. The financial analysis use case maps to focused, desktop contexts.
Tensions I Navigated
Conversational AI vs. navigational data - resolved with the two-panel approach: chat for dialogue, panel for exploration
Warm editorial identity vs. technical AI product - resolved with the later iteration: Syne font (geometric, architectural) replaces Playfair, teal replaces sage, but warmth is preserved in spacing and tone
Showing Flo’s intelligence without over-promising - the ‘Insight’ strip is labeled clearly; Flo explains its reasoning, not just its conclusions
Design Execution
I chose a conversational-first layout over a dashboard because the mental model shifts from 'reading data' to 'getting an answer' that sequence changes everything about information hierarchy.
DM Mono for all financial values creates a visual language that separates data from prose, the user always knows when they're looking at a number vs. an explanation."
The three-column split was deliberate, chat stays narrow so it never feels like a spreadsheet. The data panel only exists to deepen what Flo already said.
Expected User Outcomes
This is a concept product, so there are no live metrics. But the design communicates specific, testable hypotheses:
Expected User Outcomes
<30s
Time to Understand Expenses
+40%
Navigation Efficiency
+35%
Insight Awareness
-30%
Cognitive Load
Reflection
What I'd Do Differently
The thing I’m most uncertain about: the chat history as the primary navigation model. Having five recent conversations in the sidebar feels right for a power user who asks Flo things daily. But for someone who opens the app once a month? The sidebar could feel empty in a way that undermines the product.
I’d want to test two variants: one where the sidebar defaults to recent chats (current design), and one where it defaults to smart suggestions, Flo proactively surfacing things worth asking about before you even type anything.
What I Learned
The hardest part of designing a conversational product is resisting the urge to also make it a dashboard. They want different things from you.
The data panel was the right call, but its relationship to the chat panel took several iterations to feel natural rather than forced.
Selected Works
Flo
A concept product that makes business finances feel less like a chore. Flo is an AI-powered finance tool where you just ask questions and your numbers talk back.
M