Rabbit Portal : AI-driven Search Tool
Industry
Product Design
Team
Academic Project
Role
UX/UI Designer
Project Timeline
4 months
An AI-driven search tool which helps to bridge the gap between communities and libraries by using smart prompts to deliver enhanced, relevant search results. By promoting library programs and events, it redefines libraries as vibrant community hubs, not just places for books.
An AI-driven search tool which helps to bridge the gap between communities and libraries by using smart prompts to deliver enhanced, relevant search results. By promoting library programs and events, it redefines libraries as vibrant community hubs, not just places for books.
Background
We were presented with a design challenge by industry partners to identify an underserved sector and develop an AI tool tailored to the needs of consumers within that sector.
Project Overview
Our goal was to develop an AI-driven tool that can enhance the services provided by libraries. By leveraging the transformative power of AI, this innovative solution aims to not only address existing challenges but also capitalize on opportunities to revolutionize the library experience, ultimately reshaping the future of both the workplace and library services.
User Research
As the team’s debater and facilitator, I led our brainstorming sessions to identify an underserved sector that also aligned with our interests. We ultimately decided to focus on community-driven services—specifically libraries and community centres—and chose the Toronto Public Library as our primary source for research insights.
Background
We were presented with a design challenge by industry partners to identify an underserved sector and develop an AI tool tailored to the needs of consumers within that sector.
Project Overview
Our goal was to develop an AI-driven tool that can enhance the services provided by libraries. By leveraging the transformative power of AI, this innovative solution aims to not only address existing challenges but also capitalize on opportunities to revolutionize the library experience, ultimately reshaping the future of both the workplace and library services.
User Research
As the team’s debater and facilitator, I led our brainstorming sessions to identify an underserved sector that also aligned with our interests. We ultimately decided to focus on community-driven services—specifically libraries and community centres—and chose the Toronto Public Library as our primary source for research insights.
As we explored the Toronto Public Library website and its offerings, we discovered a wide range of services and programs that were completely new to us as a group.
As we explored the Toronto Public Library website and its offerings, we discovered a wide range of services and programs that were completely new to us as a group.
Our goal became identifying gaps in library usage and services to design AI solutions that enhance user experiences, reduce staff workload, and attract new visitors—ultimately boosting overall library efficiency and community engagement.
Our goal became identifying gaps in library usage and services to design AI solutions that enhance user experiences, reduce staff workload, and attract new visitors—ultimately boosting overall library efficiency and community engagement.
Observational Research
Our team members visited different Toronto Public Library branches to carry out on-site observations, aiming to gather insights into the library environment and user experience. I personally visited the Oakville Public Library.
Each of us visited at different times throughout the week, focusing our observations on the following key aspects:
Available facilities and services
Time of day and day of the week
Demographics of library patrons
This allowed us to collect a diverse range of data and experiences, contributing to a more comprehensive understanding of the library's operations and the needs of its users.
Observational Research
Our team members visited different Toronto Public Library branches to carry out on-site observations, aiming to gather insights into the library environment and user experience. I personally visited the Oakville Public Library.
Each of us visited at different times throughout the week, focusing our observations on the following key aspects:
Available facilities and services
Time of day and day of the week
Demographics of library patrons
This allowed us to collect a diverse range of data and experiences, contributing to a more comprehensive understanding of the library's operations and the needs of its users.
Assumptions
1) Libraries are solely for book lending. They only offer printed books for rent or buy.
2) Libraries are becoming obsolete due to the rise of digital resources, overlooking their ability to adapt and offer modern services like e-books, digital databases, and technology resources.
3) Libraries have strict quiet level and rules in place. They often serve as a place to work in silence.
4) The public's understanding of AI is fragmented, with diverse viewpoints emerging from individuals' varied comprehension levels.
5) People have differing views on the usage of AI because of its complexities and the way it’s portrayed in recent media coverage.
Assumptions
1) Libraries are solely for book lending. They only offer printed books for rent or buy.
2) Libraries are becoming obsolete due to the rise of digital resources, overlooking their ability to adapt and offer modern services like e-books, digital databases, and technology resources.
3) Libraries have strict quiet level and rules in place. They often serve as a place to work in silence.
4) The public's understanding of AI is fragmented, with diverse viewpoints emerging from individuals' varied comprehension levels.
5) People have differing views on the usage of AI because of its complexities and the way it’s portrayed in recent media coverage.
Hypothesis
1) Enhancing digital library services will result in increased library usage and more people engaging with accessible public education and resources.
2) If artificial intelligence is presented in a manner that is approachable and not overwhelming, users will be more inclined to engage with and utilize AI technology.
Hypothesis
1) Enhancing digital library services will result in increased library usage and more people engaging with accessible public education and resources.
2) If artificial intelligence is presented in a manner that is approachable and not overwhelming, users will be more inclined to engage with and utilize AI technology.
User Interviews
Our methodological approach for the research was screening questions and hybrid semi-structured online interviews .
By focusing on in-depth, face-to-face interactions, we gathered rich information and personal accounts that can be analyzed for a comprehensive understanding of the subject matter.
This method allowed us for more detailed and nuanced insights compared to survey-based research, enabling us to better empathize with and address the unique needs and perspectives of each participant.
Our participants consisted of library patrons, non goers and library staff
We conducted total of 7 interviews in which I did 3 interviews which included 2 library patrons and 1 library staff member.
User Interviews
Our methodological approach for the research was screening questions and hybrid semi-structured online interviews .
By focusing on in-depth, face-to-face interactions, we gathered rich information and personal accounts that can be analyzed for a comprehensive understanding of the subject matter.
This method allowed us for more detailed and nuanced insights compared to survey-based research, enabling us to better empathize with and address the unique needs and perspectives of each participant.
Our participants consisted of library patrons, non goers and library staff
We conducted total of 7 interviews in which I did 3 interviews which included 2 library patrons and 1 library staff member.
Interview Findings
With all the findings we came up with some points for our product.
What to avoid
Clippy
Invasive prompting of the AI tool
Total replacement of librarians
Providing uninformed answers for complex questions
Using technically advance language when interacting with users
What we want to achieve
Clear prompts for users
Engaging interactions with AI using plain language
Reference multiple knowledge sources to find trusted info.
Liberate librarians from laborious and repetitive work.
Interview Findings
With all the findings we came up with some points for our product.
What to avoid
Clippy
Invasive prompting of the AI tool
Total replacement of librarians
Providing uninformed answers for complex questions
Using technically advance language when interacting with users
What we want to achieve
Clear prompts for users
Engaging interactions with AI using plain language
Reference multiple knowledge sources to find trusted info.
Liberate librarians from laborious and repetitive work.
Design Phase 1
After considering all the findings and insights from the interviews through affinity mapping, as a group we came up with the idea of an AI chatbot. But, as we found through our interview findings that people are reluctant towards the use of AI, I came up with the idea of AI driven search tool, which not only helps the user in their search but can also create a more personalized experience while giving priority to other program and events rather than just books.
So we split up in 2 groups (2 in each) and explored how search engine and chatbot would look like.
Design Phase 1
After considering all the findings and insights from the interviews through affinity mapping, as a group we came up with the idea of an AI chatbot. But, as we found through our interview findings that people are reluctant towards the use of AI, I came up with the idea of AI driven search tool, which not only helps the user in their search but can also create a more personalized experience while giving priority to other program and events rather than just books.
So we split up in 2 groups (2 in each) and explored how search engine and chatbot would look like.
User Flow
We started with this initial user flow:
User Flow
We started with this initial user flow:
For our search engine exploration, we designed a simple yet effective feature by incorporating prompt selections to help users access relevant results more quickly. We also redesigned the search results layout to improve clarity and seamlessly integrate the prompt options.
For our search engine exploration, we designed a simple yet effective feature by incorporating prompt selections to help users access relevant results more quickly. We also redesigned the search results layout to improve clarity and seamlessly integrate the prompt options.
The chatbot followed a similar approach by using prompts, but with more detailed guidance and refined search results. Both solutions reference the "Ask a Librarian" service as a fallback when AI support falls short. Our intention was to assist users during the initial search process, not to replace the valuable role of librarians.
The chatbot followed a similar approach by using prompts, but with more detailed guidance and refined search results. Both solutions reference the "Ask a Librarian" service as a fallback when AI support falls short. Our intention was to assist users during the initial search process, not to replace the valuable role of librarians.
However, based on our findings, the chatbot felt somewhat intrusive, which we wanted to avoid. As a result, we developed a unique solution that combined an AI-driven search tool with a chatbot—where the search engine supported broad, exploratory searches and the chatbot provided more concise, targeted results.
However, based on our findings, the chatbot felt somewhat intrusive, which we wanted to avoid. As a result, we developed a unique solution that combined an AI-driven search tool with a chatbot—where the search engine supported broad, exploratory searches and the chatbot provided more concise, targeted results.
Design Phase 2
After meeting with the stakeholders, we were challenged to design a solution that could be used by libraries globally, not just within the Toronto Public Library system. After weighing the pros and cons of both the search engine and chatbot, we decided to move forward with the AI-driven search tool— and that's when Rabbit Portal came into existence!
Design Phase 2
After meeting with the stakeholders, we were challenged to design a solution that could be used by libraries globally, not just within the Toronto Public Library system. After weighing the pros and cons of both the search engine and chatbot, we decided to move forward with the AI-driven search tool— and that's when Rabbit Portal came into existence!
Through our research on search optimization in Canadian libraries, we discovered that most libraries use “Bibliocommons” for search functionality. However, the twist was that Bibliocommons primarily focuses on book-related searches and lacks features that support the promotion and engagement of other library programs and events.
Through our research on search optimization in Canadian libraries, we discovered that most libraries use “Bibliocommons” for search functionality. However, the twist was that Bibliocommons primarily focuses on book-related searches and lacks features that support the promotion and engagement of other library programs and events.
Design Style
We decided to go with a Metro design, the purpose of choosing this design style was to not promote a single category and let the user dive into the options. As metro design doesn't really focus on a particular tile or category it helped us to tackle the problem of library = books.
Design Style
We decided to go with a Metro design, the purpose of choosing this design style was to not promote a single category and let the user dive into the options. As metro design doesn't really focus on a particular tile or category it helped us to tackle the problem of library = books.
As designers we had more leverage on what to show and how to have more impact on users and search results. With this we could show other program or events on the big tiles to promote more programs and services provided by the libraries.
As designers we had more leverage on what to show and how to have more impact on users and search results. With this we could show other program or events on the big tiles to promote more programs and services provided by the libraries.
Design Challenges
We had the design, we had the product, but where exactly this product exist in this library system? Is it a plugin? A different website? How are libraries going to use this? We hit a setback at this point as we were not aligned with what this product was going to look like and where does it fit.
So I took on the role of facilitator and invited each team member to share their concerns individually. Interestingly, we all had similar questions, but one teammate had a clear idea in mind. We listened to her suggestion, discussed it as a group, and ultimately agreed to integrate our solution directly into the library’s website—specifically within the search engine itself.
We had this issue because the stakeholder mentioned that we have to develop this product for all the libraries and not particularly for the TPL. We tackled it quickly and got the whole group on the track again.
To tackle this challenge, we as a group did a design sprint to gather some ideas and it also served as a good practice to learn and see how we can implement this design in our product.
Design Challenges
We had the design, we had the product, but where exactly this product exist in this library system? Is it a plugin? A different website? How are libraries going to use this? We hit a setback at this point as we were not aligned with what this product was going to look like and where does it fit.
So I took on the role of facilitator and invited each team member to share their concerns individually. Interestingly, we all had similar questions, but one teammate had a clear idea in mind. We listened to her suggestion, discussed it as a group, and ultimately agreed to integrate our solution directly into the library’s website—specifically within the search engine itself.
We had this issue because the stakeholder mentioned that we have to develop this product for all the libraries and not particularly for the TPL. We tackled it quickly and got the whole group on the track again.
To tackle this challenge, we as a group did a design sprint to gather some ideas and it also served as a good practice to learn and see how we can implement this design in our product.
Final Design Phase
We decided to implement a dashboard within the search interface to keep the entire search experience on a single page, ensuring content remains relevant and the user stays focused. While we began developing user flows to guide our wireframes, I collaborated with a teammate to build the design system simultaneously.
Final Design Phase
We decided to implement a dashboard within the search interface to keep the entire search experience on a single page, ensuring content remains relevant and the user stays focused. While we began developing user flows to guide our wireframes, I collaborated with a teammate to build the design system simultaneously.
This was our initial iteration of Rabbit Portal, which we tested using the Burlington Public Library as our pilot case.
This was our initial iteration of Rabbit Portal, which we tested using the Burlington Public Library as our pilot case.
User Testing
We launched a test on UserTesting.com and got valuable feedback from the testers. These were the following main feedbacks from the testers:
They perceived this dashboard as a popup screen.
The text was small to read for them and multiple search bar confused them. But at the same time, the search was very helpful.
They did not had to scroll a lot to search and also the prompts helped them get the satisfying results
User Testing
We launched a test on UserTesting.com and got valuable feedback from the testers. These were the following main feedbacks from the testers:
They perceived this dashboard as a popup screen.
The text was small to read for them and multiple search bar confused them. But at the same time, the search was very helpful.
They did not had to scroll a lot to search and also the prompts helped them get the satisfying results
Final Design & Prototype
We iterated our design based on the user testing feedback like removing the dashboard and making the Rabbit Portal Search as a page itself. The font was changed and more elements were added to show the products presence.
Final Design & Prototype
We iterated our design based on the user testing feedback like removing the dashboard and making the Rabbit Portal Search as a page itself. The font was changed and more elements were added to show the products presence.
Figma Prototype Link
https://www.figma.com/proto/
Figma Prototype Link
https://www.figma.com/proto/
Key Impacts
For the consumers, the AI‑Driven Search Tool transforms the library experience by delivering context‑aware, prompt‑based discovery across books, digital archives, community programs, and events, reducing average search time by 60% while driving a 45% increase in resource utilization and program attendance.
For the business:
Administrators receive real‑time analytics on search patterns to inform acquisitions and outreach, automatically surfacing relevant workshops and services—which can boost event participation by up to 50% without extra marketing effort.
An intuitive dashboard streamlines content tagging and updates, cutting repetitive inquiry handling by 30% and freeing staff to pursue high‑impact initiatives.
Continuous machine‑learning refinement ensures ever‑improving relevance and accuracy as new materials and programs are added, positioning libraries as dynamic information hubs that seamlessly bridge community needs with library offerings.
Key Impacts
For the consumers, the AI‑Driven Search Tool transforms the library experience by delivering context‑aware, prompt‑based discovery across books, digital archives, community programs, and events, reducing average search time by 60% while driving a 45% increase in resource utilization and program attendance.
For the business:
Administrators receive real‑time analytics on search patterns to inform acquisitions and outreach, automatically surfacing relevant workshops and services—which can boost event participation by up to 50% without extra marketing effort.
An intuitive dashboard streamlines content tagging and updates, cutting repetitive inquiry handling by 30% and freeing staff to pursue high‑impact initiatives.
Continuous machine‑learning refinement ensures ever‑improving relevance and accuracy as new materials and programs are added, positioning libraries as dynamic information hubs that seamlessly bridge community needs with library offerings.
Key Learnings
Learned to conduct comprehensive user research that drives data‑informed design decisions
Mastered iterative user‑testing methods to validate hypotheses and refine interactions
Developed rapid prototypes to accelerate feedback loops and enhance feature viability
Gained proficiency in interviews to uncover user needs and align objectives
Built scalable product and design systems, understanding their role in consistency and efficiency
Cultivated cross‑functional collaboration skills to tackle challenges and deliver cohesive solutions
Key Learnings
Learned to conduct comprehensive user research that drives data‑informed design decisions
Mastered iterative user‑testing methods to validate hypotheses and refine interactions
Developed rapid prototypes to accelerate feedback loops and enhance feature viability
Gained proficiency in interviews to uncover user needs and align objectives
Built scalable product and design systems, understanding their role in consistency and efficiency
Cultivated cross‑functional collaboration skills to tackle challenges and deliver cohesive solutions
NEXT PROJECTS
