Designing AI Travel Prototype
Finished my prototype and started A/B testing to explore how XAI affects user experience.
Aug 3, 2025
Aug 4, 2025
·
4
min read
.jpg)
From Wireframes to Functional Prototype
This week, I completed the design of my interactive prototype, bringing the previously developed wireframes to life. The focus of this prototype is to explore how different Explainable AI (XAI) strategies can impact user experience in AI-generated travel planning for leisure travellers.
The design decisions were guided by insights from both literature and user interviews. Key elements like AI confidence scores, tooltips explaining suggestions, and semantic labels were included to provide contextual clarity. These features aim to reduce the black-box nature of AI and make the recommendations more transparent and meaningful to users.
A/B Testing and Follow-Up Interviews
To evaluate XAI effectiveness, I created a prototype for simple A/B testing. Participants will see two versions of the same travel plan: one without explainability (baseline) and one with XAI features. After reviewing each, they’ll complete a short Likert-scale questionnaire. A follow-up semi-structured interview will provide deeper insights into their experiences, preferences, and views on the usefulness of explanations.
.jpg)
.jpg)
Preparing for Usability Testing
The test includes five Likert-scale questions, rated on a 10-point scale (1 = Strongly Disagree, 10 = Strongly Agree):
- I understood why these travel suggestions were made.
- The travel plan felt trustworthy.
- I felt confident following this itinerary.
- The recommendations matched my interests.
- The overall experience was satisfying.
I’ve scheduled participants to test the prototype this week. Their feedback and interview insights will help shape the next phase of my thesis.
.jpg)