The goal of this project was to minimize the time users spend searching for physical books inside a library. Previously, when users reserved a book via the library app and arrived in person to pick it up, they often spent several minutes deciphering complex library blueprints and wandering through aisles to locate their selection.
The Core Question: How might we help individuals quickly locate their reserved books, minimizing confusion and streamlining the pickup process?
<aside> 💡
The percentage of book borrowers in Canada who visited the library in-person increased from 59% in 2020 to 90% in 2024;
The most popular reasons for Canadian library book borrowers to visit the library in 2024 were to pick up holds (39%).
Canadian Book Consumer Study 2024
</aside>
Book Finder is a cross-platform solution designed to help users instantly locate any physical title within a library's collection. The platform provides a clear overview of the book's details, real-time availability, and precise, step-by-step navigation. To maximize accessibility, no account registration is required; users can seamlessly access the tool via their smartphones or dedicated self-service kiosks located at the library entrance.
<aside>
Can be used in different platforms, with no need to create a profile
Shows complete information about the title and its status
Shows a map with accurate directions to your next reading
Can be integrated with book’s reservation app
</aside>
The visual concept focuses on accessibility, clarity, and minimalism. The logo utilizes clean lines and geometric shapes to represent books on a shelf, paired with a modern sans-serif typeface to ensure immediate readability.
The color palette was selected to balance utility and emotion:
.png)
To accelerate the prototyping phase, I began by establishing a robust foundation of design tokens for the components and layout. To optimize efficiency and reduce manual setup, I integrated a Large Language Model (LLM) into my engineering workflow.
I used a precise prompt detailing the structural requirements for the system. For example, I instructed the LLM to generate a custom numeric weight scale (50 to 900) for the functional grays and brand accents, formatted explicitly as a structured JSON file.