Good Samaritan Shelter
Good Samaritan Shelter is a nonprofit providing housing and support services to vulnerable individuals across six Santa Barbara County cities. Their resource-sharing system relied on a manually updated spreadsheet with 380+ records, making it time-consuming for staff to maintain and difficult for users to find current services through their newsletter. My team and I designed a mobile platform that transforms these spreadsheets into a searchable tool, enabling unhoused individuals and families to easily discover needed services.
Role
Product Designer
Timeline
12 weeks
Team
3 Product Designers
1 Product Manager
2 Engineers
My Role
I led the design of the map view, list view, and AI search features, which directly addressed the problem of clients struggling to navigate outdated spreadsheets by making services easy to discover, filter, and compare.
Impact
Set to launch in Spring 2026, the platform is projected to support over 2,000 individuals, the number Good Samaritan Shelter currently hosts across its housing and support programs yearly. In informal user testing with 8 participants, we observed a 68% reduction in the time required to find accurate service information.
Context
Over 380 services were tracked in a single spreadsheet that had to be updated by hand, leading to frequent errors and outdated information. The resource data was scattered across six different city sources, with no unified database. A small staff team was left to reconcile everything manually, making the system unsustainable at scale. Without a reliable data foundation, both staff and unhoused individuals struggled to access accurate, timely service information.
Accessible Design
Inclusive Experiences for a Wide Range of Abilities
Inclusive Design
Recognizing this barrier, we set out to design an inclusive experience that helps vulnerable individuals quickly access essential services. We established a brand identity that meets WCAG AA accessibility standards and streamlined development to ensure compatibility with screen readers.
Designing for AI
I explored how AI could enhance semantic search by intelligently mapping user queries to services across different municipal vocabularies in six Santa Barbara County cities (Goleta, Santa Maria, Lompoc, Santa Barbara, Grover Beach, and San Luis Obispo). For example, a search for 'food assistance' could connect to 'SNAP benefits' in Santa Barbara or 'emergency food pantry' in Lompoc, despite different terminology for similar services.
Stakeholder Needs
Balancing Competing Priorities
Along the way, one of my biggest challenges was reconciling different perspectives and competing priorities by designing an experience that balanced both user needs and stakeholder goals.
User Research
User Research
89% of participants rely directly on case workers for assistance
Participants often found online information outdated, inaccurate, or incomplete, leading many to bypass digital tools entirely, making it harder to find help quickly.
88% of participants prioritize location & hours
Google Maps often showed incorrect data or missed smaller local services within Santa Barbara, turning the search to find essential services into a stressful experience.
68% of participants rely on government-issued androids
Most users rely on government-issued devices with limited data and unstable internet, so we prioritized a mobile-first app designed to function across varying levels of connectivity.
Framing the Opportunity
How might we design a 0→1 accessible mobile app that helps individuals easily discover and connect to essential support services?
Ideation
Ideation
To translate our findings into design, we began by mapping the map and list view user flows. While we anticipated that the information hierarchy and content would evolve, establishing these low-fidelity flows were a critical starting point.
Design Rationale
Iterative Prototyping
Iterative Prototyping
As our prototypes evolved, we constructively critiqued each other’s ideas, advocated for our users, explored multiple variations of key app screens, and experimented with information hierarchy.
Technical Constraints
Designing for an AI Search Experience
Iterative Prototyping
As our prototypes evolved, we constructively critiqued each other’s ideas, advocated for our users, explored multiple variations of key app screens, and experimented with information hierarchy.
Reusable UI Components
Atoms, Molecules, Organisms
I contributed to a library of reusable UI components, from buttons to tooltips, with clearly defined interaction states (pressed, active, disabled) that met minimum touch-target requirements, helping streamline development while ensuring accessibility.
Offline-First Designs
Optimizing for Low-Bandwidth
Iterative Prototyping
We also treated offline design as an important accessibility and reliability challenge since many of our users experience experience unstable or limited internet connectivity.
Hi-Fi Prototype
Final Prototype
Iterative Prototyping
Putting it all together, below, I demonstrate three simple user flows for finding support, allowing users to browse details in the list view, explore options on an interactive map, and seamlessly search for services.
List View
After onboarding, users land on the List View by default, which consumes less data and enables them to quickly browse services by department or category, personalized to their location.
Map View
The Map View enables users to spatially locate nearby walk-in resources. By integrating the Google Maps SDK, users can explore locations, view details, and plan their route easily.
Expanded Service Card
The expanded services card allows users to complete the eligibility application, including information about available services and a step-by-step guide detailing how to access each specific service.
Semantic Search
Some users prefer to search directly for services. The AI-powered search dynamically surfaces the most relevant resources, helping users quickly find what they need without browsing through lists or maps.
Conclusion
Next Steps
User Research
Preparing for Developer Handoff
We prioritized creating documentation for developers that was clear and comprehensive, with a focus on accessibility, content, micro-interactions, state definitions, and specs detailing our design system.