San Diego Waste Collection Portal

CHALLENGE

San Diego residents must select new sizes for trash, recycling, and compost bins (weekly/bi-weekly). The existing portal created confusion because it lacked:

Clear guidance on which bin size fits a household
Decision support for cost and schedule tradeoffs
Helpful error/confirmation patterns during checkout
Real-time help (FAQs were limited and static)
Inclusive support for multilingual and accessibility needs

Opportunity: Replace “hunt and guess” with a guided, human-like experience using an AI assistant.

Information

TYPE

Class concept project (AI in UX/UI)

ROLE

Lead UX/UI Consultant, Visual Designer

TOOLS

Figma, ChatGPT, Perplexity, Lovable/UI prototyping tools

PROCESS

Research & Insights
I reviewed public data and UX patterns in government services to understand resident needs and accessibility gaps.
Key insight: A large portion of residents speak languages other than English at home, and users prefer real-time conversational guidance over static FAQs.

DISCOVERY

Using ChatGPT and Perplexity, I mapped likely questions and usability breakdowns in bin selection and checkout.
Patterns identified:

Guided decision-making
Transparent pricing and service explanation
Multilingual communication
Personalized assistance

AI tools used:

ChatGPT

Perplexity

Granola

Primary Research
I conducted semi-structured interviews and think-aloud walkthroughs with residents representing three user types:

Eco-Conscious Family
Busy Professional Couple
Retired Homeowner

What I learned: People needed plain-language recommendations, reassurance, and the ability to ask follow-up questions in the moment.

Define + Design
I created a journey map to pinpoint drop-off moments and designed an AI chatbot concept integrated across the portal.

AI tools used:

Make

html. to. design

Figma Make

Lovable

UIzard

UX Pilot

FINAL DELIVERY

AI-Powered Multilingual Chatbot (Portal-wide)
Core features:

Language support: English, Spanish, Tagalog, Mandarin, Vietnamese, Arabic
Voice dictation (accessibility and ease of use)
Guided bin selection based on household size and pickup frequency
Clear explanations of pricing, schedule, and “what happens next”
Escalation path to human support for complex cases

Outputs: Wireframes and mid-fidelity prototype showing entry points, key conversation flows, and translation/voice interactions.

REFLECTION

This project showed me how AI can accelerate UX work (discovery, dialogue flows, multilingual content) while requiring strong human oversight. I treated AI insights as hypotheses, validated with real user conversations, and designed with responsible AI principles: transparency, privacy-first behavior, accessibility, and human escalation.

“I redesigned a confusing civic form into a guided, multilingual AI-assisted experience that improves clarity, accessibility, and trust.”