Designing an AI-powered commuting assistant to reduce missed GO bus fares
Timeline
3 Weeks (Jul 2025)
Role
Lead Product Designer
Team
2 designers
Skills
Product thinking, user research, interaction design
Project Context
Design brief
As part of a Customer Experience Design course, we were tasked with addressing a real-world customer experience challenge by identifying pain points, evaluating constraints, and proposing a solution that delivers clear value to users.
Problem space
University students relying on GO Transit often miss their buses despite using transit apps. Unpredictable delays, inaccurate updates, and demanding academic schedules make commuting stressful and unreliable, frequently resulting in missed classes and increased daily stress.
Where the Experience Broke Down
PROBLEM DEFINITION 🎯
Current Workflow
Existing transit apps show schedules but fail to account for real-world variability like traffic, weather, and individual routines. Students are left guessing when to leave, often missing buses despite consulting multiple transit tools.
How might we ensure students never miss their GO buses and make commuting more reliable and stress-free?
Understanding the Existing Experience
PRIMARY RESEARCH 🔎
User Interviews
I conducted surveys and interviews with undergraduate students who regularly use GO Transit to understand the scope of the problem.
Key research insights
90% of surveyed students had missed a bus due to inaccurate or delayed app updates
Students reported high stress and disrupted routines from unreliable commute planning
SYNTHESIZING RESEARCH 🔎
Customer Journey Mapping
We synthesized research insights into a customer journey map to visualize the existing commute workflow and identify intervention opportunities.
Key takeaways
Account for real-world unpredictability with dynamic traffic and weather updates
Deliver personalized reminders that adapt to user habits
Use highly visible notifications (lock screen + audio) to reduce missed buses
Automate weekly bus planning based on class schedules
Designing a Clearer Path Forward
IDEATE 🖌️
User Flows
Using research insights, we mapped key user flows for the core features.
IDEATE 🖌️
Mid-Fi Wireframes
Due to the compressed timeline, we moved quickly into mid-fidelity designs while continuously sharing feedback throughout the process.
Bringing the Experience Together
PROTOTYPE ▶️
Final Design
GO Smart is a personal transit assistant that handles planning, adapts in real time, and ensures students are always informed and on time. Unlike traditional transit apps that only show schedules, GO Smart tells students exactly when to leave and why, removing guesswork and reducing missed buses.
Key features:
Learns from user behavior (e.g., consistently leaving late) and adjusts reminders to ensure on-time arrivals.
Tracks live traffic, weather, and mode of travel (walking, biking, driving) to dynamically update departure times.
Syncs with academic and personal calendars to generate a personalized weekly bus plan.
Delivers dynamic notifications paired with auditory alerts to maximize visibility.
What This Project Taught me
REFLECTIONS 💡
Lessons Learned
The power of narrowing scope
Through research and ideation, I discovered multiple pain points within the GO Transit experience. However, trying to address all of them would have significantly went out of scope. Focusing on a single, high-value problem of helping students leave on time enabled a clearer, more effective solution that directly addressed their most critical need.
Balancing feasibility with design ambition
One early concept involved allowing riders to "page" bus drivers to prevent early departures. While creative, this introduced feasibility concerns including potential overuse and driver frustration. Reframing the problem toward optimizing the passenger's workflow reinforced the importance of balancing ambition with feasibility and adapting designs when constraints arise.
Designing under time constraints requires strategic trade-offs
With only three weeks, we had to prioritize research depth, design exploration, and feature scope carefully. This taught me to identify the highest-impact areas to focus on and accept that not every aspect could be fully explored. The experience reinforced the importance of making strategic trade-offs when working under tight deadlines.
REFLECTIONS 💡
Next Steps
Conduct competitive analysis
With more time, I would analyze existing transit solutions like Citymapper, Transit app, and Google Maps to identify best practices for real-time updates, notification patterns, and predictive features. This would help refine GO Smart's approach and identify opportunities for differentiation.
Conduct usability testing with our user group
I would recruit students who commute multiple times a week to evaluate the clarity of departure recommendations, effectiveness of adaptive reminders, and comprehension of timing changes. Testing would help refine interaction flows and reduce friction in high-stress, time-sensitive moments.
Explore accessibility for notifications
Given the time-critical nature of commuting, I would evaluate how notification timing and auditory cues can be made accessible to users with hearing, visual, or attention-related needs. This includes exploring volume levels, repetition, and alternative feedback methods to ensure updates are noticeable without being overwhelming.
Validate AI predictions against real-world data
To ensure reliability, I would validate departure predictions against real-world commute data including traffic conditions, weather patterns, and historical delays. Comparing predicted versus actual outcomes would assess accuracy, identify failure cases, and inform model adjustments to increase user trust.
