A Lazy Traveler’s Travel Guide

When you discover that neither of you was the person who planned those trips in a relationship…

“You are a world-class, famous trip planner. I need your help planning a trip to Oahu and Maui. Do an extensive web search to read at least 75 to 100 different blog posts and…” he jumped into enlisting his bestie, Claude, with a quick press of the Function key on the keyboard which invoked Wispr – a medium perfect for long, detailed voice-to-text prompts and contexts for GenAI agents.

Great…just that my approach to research was different.

I needed to visualize that information on a map so I could start figuring out which cluster of places we should visit each day.

A quick Google Image search yielded helpful maps from Shaka:

But without the ability to customize it, the general guide felt more distracting than useful.

I wanted something that was visual, direct, quickly changeable, and easily shareable. I wanted to start with curated recommendations on a map, remove what didn't interest me, and reorder activities by day – like a visual Kanban board. I’d love reminders for places that need advanced reservations, too.

A tool was born.

Red Bull GenAI gives me wings.

GenAI agents have democratized coding – perfect for someone like me with limited programming experience. Finally, I could prototype quickly: sketching out general concepts I wanted. Something that would have taken me 2 weeks with manual information update now took me ~8 hours.

First with Claude:
- Imported all the research we’d done into a project for context;
- Tried out the “quick start” recommendation to build a microservice with a plan and architecture;
- Reprompted through multiple rounds to get the features functioning how I wanted.

Then with Lovable for polish:
- Prompted it to give the experience a more 'relaxed Hawaiian' aesthetic;
- When Lovable unexpectedly removed some sections, I reprompted and used the visual editor to restore them.
The result is a visual trip planner that displays recommendations on a list and an interactive map, (almost) lets me drag activities into daily schedules, and flags reservations needed.
- One big insight: Out-of-the-box results based on prompts were good to an extent, we still need trained professionals to guide the “taste-making” and giving directions to refine content hierarchy and relevancy, based on actual human expectations and behaviors.