Generative AI Product Pivots
A sequence of grounded generative-AI product pivots across legal work, document automation, and travel planning, built on reusable production capabilities.
Primary outcome
Expanded into legal tech, document AI, and travel tech; an eight-person team launched the Tripmate MVP in two weeks.
The problem
Find viable AI-native product directions while operating a small startup team and reusing a production platform originally built for EV charging and mobility.
Constraints
- A small team had to validate multiple domains without fragmenting the core platform.
- Generated answers needed grounding in domain documents or actual place data.
- Each pivot required a credible user workflow, not a standalone AI demo.
Approach and key decisions
- 01
Reframed domain workflows into focused products: D.PILOT for legal work, Documate for document automation, and Tripmate for AI travel planning and companionship.
- 02
Connected Azure OpenAI function calling with real place-search results and multi-provider map data.
- 03
Used rapid prototypes and user-validation flows to test product direction before deep platform investment.
- 04
Reused production capabilities across authentication, payments, files, chat, notifications, and deployment pipelines.
Outcomes
- Developed D.PILOT and Documate workflows for legal and document work with professional review.
- Launched a Tripmate MVP with AI itinerary planning, TravelGPT, place discovery, matching, and community features in two weeks with an eight-person team.
- Secured selection for the Microsoft Launcher program with up to KRW 500M in Azure credits.
Technology and methods
Evidence and disclosure scope
This case is an owner-authored summary. Product names, program selection, team size, and launch timing come from the portfolio resume; confidential user and business metrics are omitted.
See the full resume context