BUILDCAPSULE
PrototypeSample Project2026

AI Travel Companion

A prototype travel companion that plans around constraints — budget, energy, weather — rather than listing attractions.

Balance scale settling over a map, weighing travel options against each other

Challenge

Trip planning tools optimise for coverage. Real travellers optimise for energy and mood, and change plans mid-day.

Thinking

Treated the itinerary as a living document rather than an output: every suggestion is provisional and re-planning is the primary action, not an edge case.

Solution

A re-plannable, constraint-first itinerary prototype for mobile.

Build

Built a mobile-first prototype with a day timeline, a constraint bar, and a re-plan action that regenerates only the remaining hours while preserving anything the traveller pinned.

Experience

Re-planning animates as a fold rather than a reload — pinned items hold position while the rest reflows around them.

Outcome

Prototype demonstrates constraint-led, re-plannable itinerary design. Fictional destinations data, no bookings, no partners, no user numbers.

Reflection

Pinning turned out to be the trust mechanism; without it people would not let the model touch the plan at all.

Gallery