AI in Legal Practice
AI in Legal Practice
Qualitative user research exploring how legal practitioners interact with artificial intelligence
Qualitative user research exploring how legal practitioners interact with artificial intelligence
Overview
In early 2026, I initiated and led a qualitative research initiative to understand how legal professionals use, perceive and adopt artificial intelligence. Through eleven in-depth interviews with lawyers, conveyancers and practice directors, I explored their current use of AI, barriers to adoption, trust concerns and expectations for future tools.
The research revealed what drives trust in AI, where practitioners see its limitations and how existing workflows influence their willingness to adopt new solutions. These findings provided a clearer understanding of the opportunities and challenges shaping AI adoption across the legal industry, helping inform future product and design decisions.
In early 2026, I initiated and led a qualitative research initiative to understand how legal professionals use, perceive and adopt artificial intelligence. Through eleven in-depth interviews with lawyers, conveyancers and practice directors, I explored their current use of AI, barriers to adoption, trust concerns and expectations for future tools.
The research revealed what drives trust in AI, where practitioners see its limitations and how existing workflows influence their willingness to adopt new solutions. These findings provided a clearer understanding of the opportunities and challenges shaping AI adoption across the legal industry, helping inform future product and design decisions.
Project details
Role
Lead UX researcher and product designer
Lead UX researcher and product designer
Team
3 Product Designers
3 Product Designers
Timeline
April to May 2026
April to May 2026
Scope
11 client discovery sessions across property, family law and wills practices
11 client discovery sessions across property, family law and wills practices
The challenge
While AI adoption is accelerating across the legal sector, its use varies significantly, from administrative tasks like dictation to more complex applications such as legal drafting and contract review. I set out to understand the behaviours, attitudes and underlying barriers shaping how legal professionals adopt AI in their day-to-day work.
Rather than simply asking what features they wanted, I focused on understanding their existing workflows, where concerns around data security and accuracy limited adoption, and why human oversight remained essential in high-stakes legal work.
While AI adoption is accelerating across the legal sector, its use varies significantly, from administrative tasks like dictation to more complex applications such as legal drafting and contract review. I set out to understand the behaviours, attitudes and underlying barriers shaping how legal professionals adopt AI in their day-to-day work.
Rather than simply asking what features they wanted, I focused on understanding their existing workflows, where concerns around data security and accuracy limited adoption, and why human oversight remained essential in high-stakes legal work.
Key findings
Human review is non negotiable
Every participant confirmed that AI outputs are thoroughly reviewed before reaching clients. Fearing hallucinated case references and professional liability, lawyers treat AI as a starting point rather than a replacement for professional judgment.
Every participant confirmed that AI outputs are thoroughly reviewed before reaching clients. Fearing hallucinated case references and professional liability, lawyers treat AI as a starting point rather than a replacement for professional judgment.
AI belongs in the workflow
Legal professionals prefer AI embedded directly into their practice management software. Seamless access to matter context matters far more than raw AI capability.
Legal professionals prefer AI embedded directly into their practice management software. Seamless access to matter context matters far more than raw AI capability.
Data security is essential for adoption
Firms are cautious about using cloud-based AI tools due to concerns about sensitive client data being stored or processed overseas. Secure environments and control over where data is stored are essential requirements.
Firms are cautious about using cloud-based AI tools due to concerns about sensitive client data being stored or processed overseas. Secure environments and control over where data is stored are essential requirements.
AI struggles to match firm-specific language and precedents
AI-generated correspondence and advice letters often fail to reflect a firm's tone of voice and established legal precedents. This means practitioners, particularly senior staff, must spend extra time reviewing and editing outputs.
AI-generated correspondence and advice letters often fail to reflect a firm's tone of voice and established legal precedents. This means practitioners, particularly senior staff, must spend extra time reviewing and editing outputs.
Product opportunities
The research translated directly into four core product directions for future legal software development:
The research translated directly into four core product directions for future legal software development:
Matter intelligence and automated advice
Integrating search results directly into automated advisory summaries to save hours of manual drafting.
Integrating search results directly into automated advisory summaries to save hours of manual drafting.
Proactive matter monitoring
Shifting from reactive prompt responses to daily digests highlighting critical dates, missing searches, and pending verifications.
Shifting from reactive prompt responses to daily digests highlighting critical dates, missing searches, and pending verifications.
Client explainer layers
Providing structured, plain language outputs to help end clients understand complex search results and reduce inbound phone call volumes.
Providing structured, plain language outputs to help end clients understand complex search results and reduce inbound phone call volumes.
Outcomes and learnings
By turning raw interview notes into a structured research report, I gave product and engineering teams a clearer understanding of what drives trust and adoption of AI in legal workflows. The findings reinforced that AI capabilities alone are not enough. Successful enterprise AI solutions must prioritise data security, firm-specific context and human oversight.
Leading the research from start to finish also strengthened my ability to navigate complex and rapidly changing problem spaces, turning qualitative user insights into actionable findings that could guide future product decisions.
By turning raw interview notes into a structured research report, I gave product and engineering teams a clearer understanding of what drives trust and adoption of AI in legal workflows. The findings reinforced that AI capabilities alone are not enough. Successful enterprise AI solutions must prioritise data security, firm-specific context and human oversight.
Leading the research from start to finish also strengthened my ability to navigate complex and rapidly changing problem spaces, turning qualitative user insights into actionable findings that could guide future product decisions.