Semantic Search
Adding an AI search feature
Led the integration of semantic search, a suite of solutions providing a seamless experience across multiple applications. Initially, I focused on one application and then extended the feature to all the applications on the platform while maintaining a consistent UX.
Introduction
Integrating semantic search was a transformative project aimed at enhancing the user experience by leveraging AI to provide more accurate and relevant search results. This project not only improved the search functionality within the investigations app but also extended the feature across all the applications on the platform, maintaining a consistent UX throughout.
The Project Challenge
Shifting Focus
Initial Misalignment: Upon joining the team, the primary discussion revolved around how to display semantic search results rather than focusing on the user's search process.
Refocusing Efforts: I redirected the conversation towards understanding user requirements and optimizing the search process to ensure users could efficiently obtain relevant results.
Consistency Across Apps:
Uniform Implementation: Ensuring that the semantic search designs could be consistently implemented across all applications on the platform while maintaining a uniform user experience.
The Team
Project Managers
Lead Product Designer (Me)
Head of AI Strategy & Operations
Head of Design
Engineers
Architects
Chief Technical Officer
Engineering Founder & Chief Scientist
Personas
We focus on these personas to align the design efforts with the needs and behaviors of the target users, ensuring a user-centered approach throughout the project.
Detective:
Needs: Quick access to relevant case data.
Behavior: Uses the platform for investigative purposes, seeking detailed and accurate information.
Corporate Investigator:
Needs: Efficiently sift through large volumes of data.
Behavior: Focuses on identifying patterns and evidence in corporate data.
User Flow
The following user flow illustrates the step-by-step journey our users take, highlighting key interactions and touchpoints throughout the experience.
Components and functionality
Search widget
The user has the option to search using text, an image from the case, or to upload an image.
Text search
For semantic search, more text yields better results. To encourage users to input more text, I implemented a ticker that changes color based on the text amount.
Search results
The results listing is sorted by relevancy and when using the semantic search feature, a new column is added with the relevancy score.
Similar items
When the user find items of interest, they may want to find all the items in the case that are similar. They may also want to find items with similar text on it.
Consistency Across All Apps
Semantic search will be integrated into all Nuix Neo apps. To maintain consistency, I designed this widget to be implemented alongside search inputs in each application.
Prototype
In this prototype, you can see how the user interacts with all aspects of the Semantic Search feature. I designed the user interface using Nomi, the Nuix design system, to ensure a consistent and seamless user experience.
Research indicated that it’s important to highlight areas involving AI. Thus, I chose to use purple along with sparks, as this visual distinction helps users easily identify and trust the AI-driven components of the feature.
What People Are Saying
“I love the simplicity of how the Neo Platform has been pulled together. The navigation provides an easy and intuitive way to navigate the platform.”
— Stephen Stewart – Field CTO
“With all the noise about GenAI, the use of color throughout the UI makes it easy to tell when GenAI has been used.”
— Nuix Customer – XLR8 Conference
Conclusion
The integration of semantic search represents a significant advancement in providing a seamless and intuitive user experience across multiple applications. By shifting the focus from result display to optimizing the user’s search process, I ensured that the semantic search feature meets user needs effectively. The design of the search widget, which supports text and image inputs, enhances search flexibility and accuracy.
The implementation of a color-changing ticker encourages users to input more text, thereby improving search results. Consistency across all Nuix Neo apps was maintained by designing the widget to be integrated alongside search inputs in each application. The positive feedback from users highlights the success of our efforts in creating a simple, intuitive, and effective search experience.
This project underscores the importance of a user-centered approach, continuous collaboration, and maintaining a uniform UX across all applications. It showcases my ability to lead complex integrations, refocus project goals, and deliver designs that meet both user and business needs.
Future Considerations
As we continue to innovate, we plan to introduce additional features and enhancements to further improve the user experience and expand our market reach.