Skip to content
0→1 build·concept

TailorTrip

A travel planner that designs the trip around how you actually like to travel.

A designed-but-not-built concept. Included here honestly — to show how I frame a product before committing engineering, not to claim something that ships.

Role
Concept — product design & spec
Timeline
Concept exploration
Year
2025
Domain
AI · Travel
Problem framingConcept designCompetitive teardown
This one is a concept, not a product. It's here because how you think before you build is itself a PM skill worth showing.

The problem worth solving

Itinerary tools optimize the wrong variable.

Most AI travel tools optimize for coverage — see the most things. Experienced travelers optimize for fit — the right pace, the right density, the trip that matches how they actually like to move through a place. A planner that ignores taste produces an itinerary nobody follows by day two.

TailorTrip's premise is that the planner's real job is eliciting and respecting preference, not retrieving attractions. The hard product question is how to learn taste fast without a tedious questionnaire.

If I built it

The riskiest assumptions I'd test first

  • Can we infer travel taste from 3–4 light signals instead of a long quiz? (Biggest risk — test with a Wizard-of-Oz planner.)
  • Do people trust an AI-built itinerary enough to actually book from it?
  • Is 'fit' a strong enough wedge against incumbents that already own discovery?

Why it's still a concept

Honest call

I haven't built it. The riskiest assumption — that taste can be inferred cheaply — isn't validated, and I'd want a Wizard-of-Oz test before writing code. Shipping the honesty here matters more than padding a portfolio.

Next · case study

Netflix: the night completion rate fell

A root-cause analysis of a metric that dropped without a deploy to blame.

Continue