Overview
App store reviews showed the travel portal was Chase's top complaint. 21% of negative reviews mentioned travel usability. The work translated user pain into a prioritized roadmap tied directly to revenue impact.
The strategic insight was that users abandoned bookings not because they couldn't find what they needed, but because they couldn't verify value. This wasn't a navigation problem. It was a verification problem. A trust problem.
The solution focused on real-time conversion visibility and simplified checkout — removing uncertainty, not just friction.
The Problem
Travel rewards differentiate Chase's premium cards (Sapphire Reserve). But low redemption rates meant users were not experiencing the value they paid for. The portal could not be redesigned wholesale. Changes had to be surgical and tied to measurable outcomes.
The constraint: Not design resources. The constraint was identifying which friction points actually drove drop-off.
The Approach
AI-assisted analysis surfaced specific friction points from 21% negative review volume:
- Data-informed problem identification — Python, Pandas, NLP analysis of app store reviews
- Friction point prioritization — Identify which issues drove booking abandonment
- Trust-centered solution design — Real-time conversion, simplified checkout
- Revenue-tied roadmap — Connect UX improvements to Chase's partner booking revenue
Key insight: Users could not verify whether they were getting a good deal. Filtering was broken, but the real issue was trust. They abandoned bookings because they could not confirm reward value matched expectations. Many booked elsewhere to avoid uncertainty.
What Was Built
- AI-assisted review analysis — Python, Pandas, NLP processing of app store feedback
- Friction point prioritization framework — Identifying which issues drove actual drop-off
- Trust restoration design — Real-time point value conversion before commitment
- Simplified checkout — Removed unnecessary confirmation steps
- Revenue-tied roadmap — UX improvements connected to Chase's revenue model
Outcomes
Projected 5% booking lift from UX improvements. Design and validation completed in 10 weeks.
Why This Matters
This project demonstrates:
- Product strategy — Identifying the real constraint (trust, not navigation)
- Systems thinking — Understanding that the symptom (abandonment) pointed to a structural issue (verification)
- AI-assisted synthesis — Using NLP to extract patterns from unstructured feedback
- Conversational UX principles — Applied to non-conversational interface: trust through transparency, clarity over options
The insight generalizes: Many "UX problems" are actually trust problems. The solution is often making information visible, not adding features.