The habit-app market competes on features. We competed on personality, designing blunt, funny notification copy as the core product, then building the brand, UI and business model around that single idea. I made every design decision under a hard 54-hour deadline, and the judges' feedback validated the direction.
A startup in a weekend
The constraint isn't a side note: it shaped every design decision that follows.
Startup Weekend is Trinity's Entrepreneurial Society's flagship three day event: strangers form teams on Friday night, build a business over the weekend, and pitch to a panel of venture capitalists on Sunday.
With no time for polished research or iteration cycles, my job was to make the fewest, highest leverage design bets, and make each one legible to a panel of investors in a ten minute pitch.
The team: Surpreet (product) · Daniele (finance) · Rahul (design + engineering)
Reminders are easy to swipe away
Nobody ignores a friend who texts "get up, you said you'd go." Everybody ignores a notification that says "Reminder: Gym."
Busy young adults don't fail to build habits because they lack reminder apps: they already have Google Keep, Todoist and a native Reminders app. They fail because those reminders are emotionally weightless. A grey notification that says "Gym · 7:30am" costs nothing to dismiss.
So we reframed the problem. Instead of competing with productivity tools on organisation, FrAnK-U competes on accountability: an app that talks to you the way a blunt friend does. Hence the name, "Frank". The AI angle was in generating that persuasive, personal copy per user, rather than canned reminder text.
Three design bets, made fast
With 54 hours I couldn't explore broadly. I had to commit. These are the three bets I made, why I made them, and what each one cost.
Reminder: Gym
You have 1 task due today
Get off your ass. Time for GYM. 🏋️
Third snooze, huh? Bold strategy.
Tone of voice is the product, not a garnish
The differentiator lives in the notification, not the app. So I designed the notification copy first, and treated the app as the place you configure Frank, not the place you live.
A "nagging interval" the user sets themselves
Persistence is only charming when it's consented to. I designed a nagging interval control into the profile: you tell Frank how many times to hassle you before he gives up. It turns the app's most annoying behaviour into its most personal setting.
Ruthlessly minimal capture: two fields, one button
Todoist wins on power features; we couldn't out-feature it in a weekend and shouldn't try. Creating a reminder in FrAnK-U asks exactly two things, what and when, because our users are people who abandon complex tools, not people who crave them.
13 screens, one voice
I designed 13 screens in Figma across onboarding, capture and notifications, plus the marketing site. The visual language (a single loud coral, an emoji as the logo) was chosen to feel cheap, cheeky and confident, like the copy. Real type hierarchy, accessible contrast, and the escalation mechanic made visible. The six screens below are the ones that carry the argument: the daily loop, and the nag itself.
frank·u
The friend who won't shut up until you've done it.
Morning, Rahul.
You promised 3 things today. Frank remembers.
Today
Done · Frank approves
When
Repeat
Frank will say↻ reroll
"Get off your ass. Gym. You know the deal. ✌️"
The daily loop. The splash states the promise in eight words, no stock illustration needed. Home frames tasks as promises, with the overdue card escalating visually: tinted background, "You're late" badge, and a nag counter. Capture stays two decisions, what and when, with a live preview of what Frank will actually say. Go on, tick something off, or reroll Frank's line.
How annoying should Frank be?
You can change this any time. Frank won't take it personally.
"One polite reminder. Then I'll mind my business."
"I'll keep showing up until you cave. Like cardio for your conscience."
"I will ruin your notifications. Lovingly."
Nag 3 of 5
Get off your ass. The gym misses you.
Still here.
So is the gym. Funny how that works.
Gym time ✌️
7:30, like you asked. You got this.
Frank is speechless.
Everything's done. Enjoy the silence. It won't last.
longest yet · keep Frank quiet
The nag, dialled and delivered. The nagging interval from Bet 02 grew into "How annoying should Frank be?", with each level speaking in Frank's voice. Tap a level, Frank adjusts. The lock screen shows the escalation in one glance: three notifications, rising bluntness. And the empty state now has a job, celebrating the streak instead of saying "Nothing to show!!".
How 54 hours actually went
I worked as designer and engineer on a three person team, so the process below is about sequencing scarce hours, not a textbook double diamond.
Pitch, form, reframe
Team formed around the habit motivation idea. We narrowed "AI motivation app" down to the one sharp claim we could defend: tone of voice as the product.
Validate & position
Mapped the competitive field (Keep, Todoist, native Reminders) and picked the gap none of them will touch: personality.
Design the notification first
Wrote notification copy before drawing screens. Once the voice worked, the UI had a spec to obey.
Screens & brand
13 screens in Figma: onboarding, capture, list, profile, notification states. Coral + mono type locked in.
Landing page & deck
Built the marketing site and merch concept to carry the business model story; assembled the design sections of the pitch.
Pitch to VCs
Presented FrAnK-U to the panel. Design carried the differentiation argument.
- Social validation through social media integration
- Constant, no BS reminders
- Social validation
- Track progress
- Building good habits
- Accountability
- Get shit done
- Improve social image
- Get shit done
- Reinforcing behaviour through social validation
- Inconsistency
- Lack of motivation
- Bad habits
- Struggling to build good habits
Surpreet Patheja
Product managerDaniele Lucivero
Finance and accountingRahul Ramesh
Engineer and designerThe design argument, made financial. The value proposition canvas tied Frank's bluntness to real pains (inconsistency, weightless reminders), and the ask put numbers behind the plan: 250k, with app development and guerrilla marketing (stickers and graffiti, very Frank) taking the biggest cuts.
Outcome & honest reflection
Being honest: this was a concept sprint, not a shipped product. There are no retention numbers, and the designs never met a real user. What this project does show is the thing hackathons uniquely test: judgment under constraint. What to design first, what to cut, and how to make design decisions carry a business argument.
Design the moment of truth first
Starting with the notification, not the home screen, meant every screen after it had a purpose. I now start every project by asking where the product actually earns its keep.
A strong voice is a design system
Once "blunt friend" was defined, decisions got faster: empty states, button labels, even the merch designed themselves. Personality scaled better than any component library could in 54 hours.
Pitching is a design skill
VCs don't review Figma files; they buy narratives. Presenting to a panel taught me to frame design choices as business bets, a habit I've kept for every design review since.
What I'd do differently
I'd spend two of those 54 hours testing the notification copy on strangers at the venue. The tone was our biggest bet, and we defended it with conviction when we could have defended it with evidence.