[ 02 · Design exercise · Bending Spoons application · 2026 ]
TripUp
A group-travel app I designed and prototyped for a Bending Spoons exercise.
TripUp is a group-travel app I designed as an exercise for the Bending Spoons application process, then built as a clickable prototype. The concept is mine and every number in it is fictional.
- Role
- Product Designer
- Year
- 2026
- Tags
- Product · Prototype · Exercise
- Links
- Live prototype ↗
[ 01 · Context ]
Five friends in Lisbon, day four, and nobody has said where dinner is. The plan lives in a group chat, the shortlist in someone’s saved places, the money in a third app. Each app does its job, and the dinner still does not get picked.
There is no client and no existing product behind it. The brief was an application exercise in August 2026; the concept, screens and prototype are mine. TripUp AI is scripted inside the prototype, not connected to a model.
[ 02 · Role ]
Product designer, sole author. The Figma file and the prototype are both mine. The spec on the cover is the whole constraint list: iPhone 16 at 393×852, Figtree, one accent, Sunset #FF5A45.
I built the prototype because a decision flow is easier to judge by tapping through it than by reading about it.
[ 03 · Problem ]
A group chat is a bad instrument for deciding. Options scroll away, nobody can see who has answered, and the one person who cares ends up choosing for everyone.
The money is worse. The evening gets recorded twice, once as a plan and once as a receipt in another app, and the two never meet.
[ 04 · Decisions ]
- 01
A poll with a deadline and a destination
Every poll carries three settings you see before sending: ends in 30 minutes, votes visible live, winner goes into the 20:45 slot. I gave up the secret ballot for that. The prototype shows ‘Tasca do Vintém leads’ before you have voted, so late voters can follow the leader. I accepted that because a hidden vote with no closing time drifts the way the group chat does. You can change your vote until close.
- 02
Both the swipe deck and the poll
The AI results screen ends with two doors, ‘Start a group match’ and ‘Make it a poll instead’. Swiping suits three places nobody knows yet; a poll suits three everyone has already named. One mechanic would have been simpler to learn and half the screens to build. Instead there are two empty states, two result screens, and a Swipe tab that reads ‘Nothing to swipe yet’ most of the time.
- 03
Money stays inside the trip
Split is a tab in the same bottom bar. When the poll resolves, the plan offers ‘Add the Maré Alta bill’, and Balances keeps Settle up shut until tonight’s dinner is in. The price is a five-item bar on a 393pt screen, and a Split thinner than a dedicated tool: every expense divides evenly five ways and shows its math. Uneven splits are out of scope.
[ 05 · Outcome ]
The prototype runs end to end in a browser. An undecided 20:45 slot becomes a poll with three options and a deadline, four taps later my vote is in, and when the poll closes the slot reads Maré Alta, Won 4-1-1, with the bill one tap away.
There is no shipped version and no users. The data is fictional, the AI is scripted, and the work is dated August 2026. It is a concept I took far enough to click through, and nobody has tested it with real travellers.
[ 06 · Credits ]
- Product Designer — sole author. Exercise for the Bending Spoons application process, August 2026.
- Figma (design) · prototype hosted on Netlify
- Figtree · Sunset #FF5A45 · iPhone 16, 393×852
[ Contact ]
Let’s talk
Open to product design roles, and to good collaborations.