AI Feature
AI-Powered Chat for Investigations
Designed and led the integration of an AI-driven chat feature, enabling investigators to query data seamlessly. See how I tackled technical constraints, UX challenges, and trust concerns to enhance search efficiency.
Introduction
In this product, users needed a faster, more intuitive way to query complex datasets. I led the design and integration of an AI-driven chat feature, enabling seamless natural language search. This case study breaks down the process, challenges, and key outcomes.
Goal
The goal is to enhance usability and efficiency by integrating an AI-powered chat that enables users to interact with their data through natural language queries.
Challenge
The challenge is to leverage AI to enhance the search experience while presenting results in a clear and intuitive way.
The Team
Project Managers
Lead Product Designer (Me)
Chief Technical Officer
Engineering Founder & Chief Scientist
Head of AI Strategy & Operations
Head of Design
Engineers
Architects
As the lead designer, I spearheaded the design process and facilitated collaboration between product, engineering, and stakeholders.
My Role
My responsibilities included:
Leading discovery sessions with stakeholders
Defining user needs and pain points
Creating wireframes and prototypes
Collaborating with engineering to ensure seamless implementation
Planning usability testing for Beta users
The Problem
Investigators and analysts using Nuix Investigate often struggled with complex query-building processes. Searching through vast amounts of ingested data required significant time and expertise, limiting efficiency and accessibility for non-technical users.
The Solution
We designed an AI-powered chat feature that enables users to:
Ask natural language questions about their data
Receive summarized insights without writing complex queries
Get recommendations for refining searches
Maintain transparency with AI-generated responses linked to data sources
My Design Approach
I collaborated with the CTO, Head of AI, development architects, and product managers to define key requirements for integrating AI chat within Nuix Investigate.
Requirement Gathering
I conducted research on how Open Web AI could be integrated into Nuix Neo interface while maintaining a frictionless user experience.
Integration Research
Based on the research findings, I developed multiple design options and presented them to engineering, product, and AI teams.
Design & Iterate
Since there was no time for upfront discovery and usability testing, I developed a structured beta testing plan to gather post-release feedback.
User Testing Plan
User Flow
User Testing Plan
Prepare Testing Materials
Define key test scenarios.
Develop a survey for accuracy, usability, and trust.
Create user testing guidelines.
Execute Test Scenarios
Basic Search: Ask AI chat natural language questions.
Complex Queries: Refine search using filters and follow-ups.
Result Interpretation: Assess AI-generated summaries and sources.
Error Handling: Evaluate AI responses to ambiguous queries.
Collect User Feedback
Survey & Ratings: Evaluate AI accuracy, clarity, and usefulness.
Task Time: Compare AI vs. traditional search efficiency.
Session Observations: Identify usability friction points.
Open Feedback: Gather user frustrations and suggestions.
Identify key usability and accuracy issues.
Measure adoption, trust, and efficiency gains.
Prioritize fixes and enhancements.
Analyze & Prioritize Improvements
Conclusion
By integrating an AI-powered chat, we transformed how users interact with their data. This project not only improved efficiency but also laid the foundation for further AI-driven enhancements in investigative workflows. However, beyond the feature itself, the design process revealed key learnings about rapid iteration, trust-building in AI interactions, and the value of early collaboration.
Working across teams—from C-level leadership to engineers and product managers—was crucial in aligning technical feasibility with user needs. Engaging with stakeholders at every stage ensured that design was not just an afterthought but a strategic driver of the feature’s success. Despite time constraints, the iterative approach allowed us to refine functionality, improve the user experience, and address concerns about accuracy and transparency.
Take Aways
AI-driven features require clear trust mechanisms to be effective. Involve design from the start of feature development, especially when time is tight.
Cross-functional collaboration accelerates adoption and alignment. Close partnerships across disciplines ensured AI was seamlessly integrated into Nuix Neo, delivering an intuitive and powerful search experience.
Iterative testing and feedback are critical. Even with limited discovery time, structured beta testing provided valuable insights to refine the final product.