Simplifying Natural Language Search for Improved User Experience  

User Interviews • User Flows • Wireframes • Prototyping • Usability Testing

Summary

AnswerRocket is an AI-driven data analytics platform that enables users to ask natural language questions about their data and interact with the results. In this project, we employed user research and prototyping to identify key challenges data analysts faced when searching and analyzing their data. Our goal was to explore more effective solutions to improve this process.

Outcome

This baseline user research project ended up being the foundational project for Max, AnswerRocket's key product offering contributing to over $19 million in sales annually. Read more about that project here

Development Outpacing Design

A key challenge for the product team was that development consistently outpaced design, necessitating the creation of a design process that effectively integrated with the sprint cycle.

Collaborated with the product and development teams to incorporate user research and upfront design time into the sprint cycles.

Collaborating with Services Team to Align and Engage Target Customers

I collaborated with the services team to identify and engage customers aligned with our target personas for user interviews, while also reviewing customer interviews from the past year. This research revealed several key insights:

• Need a data glossary/dictionary for metrics and dimensions.
• Require references for constructing queries and hiding incompatible keywords.
• Desire more tutorials.
• Need a "cheat sheet".
• Lack of awareness about data changes, updates, or missing data.
• Prefer a more conversational interface, similar to Google.
• Difficulty asking questions in the exact needed format.
• Unfamiliarity with the data.
• Lack of awareness about available data and load timing .
• Uncertainty about data formatting.

Guided Task Completion

I explored the concept of incorporating guided task completion and onboarding to enhance users' familiarity with their data and search functionality.

This exploration led to the development of 'Missions'.

Missions are structured mini-onboarding sessions designed to teach users about specific, complex features of the application through guided tasks. Our aim was to employ this approach to effectively guide users through the process.

Missions were born from this idea of Guided Task Completion

Conveying our vision for Missions to stakeholders

The user flows were instrumental in conveying our vision for Missions to stakeholders. Based on insights from user interviews, we recognized the value of educating users on optimal navigation within AnswerRocket.

Using design user stories and userflows to communicate the ideas to stakeholders.

Wireframes

We opted to use low-fidelity wireframes to facilitate early feedback and iterative improvements before ever finalizing the design. By testing these wireframes with 22 users, we aimed to gain initial insights and secure stakeholder approval to proceed with the design project. This early feedback was crucial for validating the concept. Below you are seeing some of the Mission wireframes.

I chose to use wireframes to test the concepts for Missions, allowing us to validate the idea without committing extensive engineering resources.

Remote Usability Testing

We were progressing rapidly and determined that MAZE remote usability testing would be the most effective approach. We set up a low-fidelity prototype and implemented a plan to gather user feedback on our proposed solution. View the Prototype

Results