JDD Lab x AI Chatbot
UX Research
UX Research for Legal AI Chatbot
With most Canadians facing legal problems within a three-year period and less than half resolving them in that time, the need for accessible legal support is critical. However, there is a disconnect between people is looking for legal help and the local legal resources and services.
Rbot is an AI-driven legal chatbot designed to bridge the gap between local legal support services and those seeking help. Rbot provides reliable, understandable, and timely legal information, ensuring everyone has access to the support they need.
ROLE
User Experience Researcher
TIMELINE
3 Months | Jan - Mar 2024
TEAM
1 × Product Manager, 1 × Visual Designers, 2 × Developers, 1 × Data Engineer
Industry
Legal
METHODOLOGY
Survey
Interview
Usability Testing
Workshop
STUDY OVERVIEW.
This generative research project I led at the Justice Data and Design Lab (JDD Lab) aimed to understand the attitudes of legal support seekers towards AI, leading to the development of a legal AI chatbot (Rbot) from scratch. The goal was to connect British Columbia (BC) residents with local legal services, ensuring that legal AI Rbot is both understandable and trustworthy.
RESEARCH OBJECTIVES.
Understand the needs, behaviors, and pain points residents face in accessing BC legal resources online.
Identify features and patterns for an AI chatbot to address BC residents' legal needs, guiding product and design decisions.
Understand the needs, behaviors, and pain points residents face in accessing BC legal resources online.
Identify features and patterns for an AI chatbot to address BC residents' legal needs, guiding product and design decisions.
Understand the needs, behaviors, and pain points residents face in accessing BC legal resources online.
Identify features and patterns for an AI chatbot to address BC residents' legal needs, guiding product and design decisions.
IMPACTS.
Strategic Impact: Pivoted product strategy based on in-depth research insights, driving alignment with user needs and business goals.
Stakeholder Collaboration Impact: Developed innovative workshop, collaborating with a product manager, developers, and a visual designer to refine research goals and deliver actionable insights for AI chatbot development.
Product Impact: Facilitated seamless integration of research feedback into prototypes and final products, enabling designers and developers to iterate rapidly in an Agile environment.
Strategic Impact: Pivoted product strategy based on in-depth research insights, driving alignment with user needs and business goals.
Stakeholder Collaboration Impact: Developed innovative workshop, collaborating with a product manager, developers, and a visual designer to refine research goals and deliver actionable insights for AI chatbot development.
Product Impact: Facilitated seamless integration of research feedback into prototypes and final products, enabling designers and developers to iterate rapidly in an Agile environment.
Strategic Impact: Pivoted product strategy based on in-depth research insights, driving alignment with user needs and business goals.
Stakeholder Collaboration Impact: Developed innovative workshop, collaborating with a product manager, developers, and a visual designer to refine research goals and deliver actionable insights for AI chatbot development.
Product Impact: Facilitated seamless integration of research feedback into prototypes and final products, enabling designers and developers to iterate rapidly in an Agile environment.
DESIGN THINKING .
I use the Design Thinking framework to guide my process—centering real user needs through empathy, defining clear problems, and supporting ideation, prototyping, and testing. This approach ensures research insights directly inform meaningful, user-driven design decisions.
I use the Design Thinking framework to guide my process—centering real user needs through empathy, defining clear problems, and supporting ideation, prototyping, and testing. This approach ensures research insights directly inform meaningful, user-driven design decisions.
I use the Design Thinking framework to guide my process—centering real user needs through empathy, defining clear problems, and supporting ideation, prototyping, and testing. This approach ensures research insights directly inform meaningful, user-driven design decisions.
HMW Question.
How might we leverage an AI chatbot to become more reliable, ensuring that people who does not engage with the formal civil justice system can access the local legal supports they need, regardless of their tech skills or legal expertise ?
How might we leverage an AI chatbot to become more reliable, ensuring that people who does not engage with the formal civil justice system can access the local legal supports they need, regardless of their tech skills or legal expertise ?
How might we leverage an AI chatbot to become more reliable, ensuring that people who does not engage with the formal civil justice system can access the local legal supports they need, regardless of their tech skills or legal expertise ?
the solution.
PROCESS.
Translating Research Insight into Design Decisions
Understanding the business problems
As a first step, it was critical to gather internal and external data that would help frame the problem, contextualize the landscape, and garner a thorough understanding of stakeholders and factors influencing the challenge ahead.
Desktop Research - ‘Access to Justice’ Landscape & technology audit
Reviewing existing research and collecting useful documentation that help the team to better understand the ‘access to justice’ issue.
Competitive Analysis
Analyzing comparative experiences provided us with an understanding of what was potentially viable and fostered our ability to have early informed conversations with stakeholders.
Informational Meeting with Stakeholders
Domain Expert (Law): Interviewing a lawyer to gather insights into existing pain points on ‘access to justice’ issue.
Domain Expert (AI): Interviewing a Data Scientist to understand the Bot's training content and how to integrate the existing product Research Engine.
Questionnaire
Conducting online surveys targeting a wide range of BC residents to gather broad insights into their legal needs and openness to using an AI-based legal advice tool; and screening potential users for the user interview.
Recruiting Criteria: to ensure a representative sample, include a balanced gender representation, only recruit residents of British Columbia, and select individuals who have dealt with various types of legal issues.
Defining the Project Strategy
User Interviews
Conducted semi-structured interviews with a diverse group of participants, including those who have faced legal issues, to dive deeper into their experiences, needs, and concerns.
User Personas
Created user personas based on users’ background information, behaviour needs and legal problem types; and bucketing them into categories, which helped the team to develop personalized user journeys that satisfy different types of users.
Facilitating Workshops
Balance between users needs and wants with how do we monetize this so we can continue reinvesting the capital to make this product even better.
Translating Research Insight into Design Decisions
Understanding the business problems
As a first step, it was critical to gather internal and external data that would help frame the problem, contextualize the landscape, and garner a thorough understanding of stakeholders and factors influencing the challenge ahead.
Desktop Research - ‘Access to Justice’ Landscape & technology audit
Reviewing existing research and collecting useful documentation that help the team to better understand the ‘access to justice’ issue.
Competitive Analysis
Analyzing comparative experiences provided us with an understanding of what was potentially viable and fostered our ability to have early informed conversations with stakeholders.
Informational Meeting with Stakeholders
Domain Expert (Law): Interviewing a lawyer to gather insights into existing pain points on ‘access to justice’ issue.
Domain Expert (AI): Interviewing a Data Scientist to understand the Bot's training content and how to integrate the existing product Research Engine.
Questionnaire
Conducting online surveys targeting a wide range of BC residents to gather broad insights into their legal needs and openness to using an AI-based legal advice tool; and screening potential users for the user interview.
Recruiting Criteria: to ensure a representative sample, include a balanced gender representation, only recruit residents of British Columbia, and select individuals who have dealt with various types of legal issues.
Defining the Project Strategy
User Interviews
Conducted semi-structured interviews with a diverse group of participants, including those who have faced legal issues, to dive deeper into their experiences, needs, and concerns.
User Personas
Created user personas based on users’ background information, behaviour needs and legal problem types; and bucketing them into categories, which helped the team to develop personalized user journeys that satisfy different types of users.
Facilitating Workshops
Balance between users needs and wants with how do we monetize this so we can continue reinvesting the capital to make this product even better.
Translating Research Insight into Design Decisions
Understanding the business problems
As a first step, it was critical to gather internal and external data that would help frame the problem, contextualize the landscape, and garner a thorough understanding of stakeholders and factors influencing the challenge ahead.
Desktop Research - ‘Access to Justice’ Landscape & technology audit
Reviewing existing research and collecting useful documentation that help the team to better understand the ‘access to justice’ issue.
Competitive Analysis
Analyzing comparative experiences provided us with an understanding of what was potentially viable and fostered our ability to have early informed conversations with stakeholders.
Informational Meeting with Stakeholders
Domain Expert (Law): Interviewing a lawyer to gather insights into existing pain points on ‘access to justice’ issue.
Domain Expert (AI): Interviewing a Data Scientist to understand the Bot's training content and how to integrate the existing product Research Engine.
Questionnaire
Conducting online surveys targeting a wide range of BC residents to gather broad insights into their legal needs and openness to using an AI-based legal advice tool; and screening potential users for the user interview.
Recruiting Criteria: to ensure a representative sample, include a balanced gender representation, only recruit residents of British Columbia, and select individuals who have dealt with various types of legal issues.
Defining the Project Strategy
User Interviews
Conducted semi-structured interviews with a diverse group of participants, including those who have faced legal issues, to dive deeper into their experiences, needs, and concerns.
User Personas
Created user personas based on users’ background information, behaviour needs and legal problem types; and bucketing them into categories, which helped the team to develop personalized user journeys that satisfy different types of users.
Facilitating Workshops
Balance between users needs and wants with how do we monetize this so we can continue reinvesting the capital to make this product even better.
MY LEARNINGS.
MY LEARNINGS.
MY LEARNINGS.
Starting strategic projects with a pragmatic approach.
Don’t start fire! You need a strategy to deliver “bad news” - and it takes practice.
Individual 1:1s with Leads before the team-wide research readouts help getting buy-in.
The more you engage stakeholders in every stage of the research process, the more they are willing to act upon your findings.
More projects to explore.
