Overview

Fuel Cycle’s Research Engine processes vast amounts of unstructured qualitative and quantitative data. However, managing and analyzing this data efficiently posed a challenge. Traditional methods required extensive manual effort, leading to delays in extracting insights.

With advancements in AI, we leveraged language models to infer objectives, generate contextual summaries, and create structured tagging systems. The AI-powered tags feature was designed to automate qualitative research analysis, reducing processing time and enhancing data organization and usability.

‍

Challenges

• Data Organization: Unstructured qualitative data lacked context, making analysis time-consuming.
• Manual Tagging Limitations: Researchers spent significant time manually tagging and categorizing qualitative responses.
• Scalability: As data volume increased, a more scalable approach to qualitative research was required.
• Adoption & Usability: Ensuring AI-generated tags and summaries were accurate, transparent, and customizable was critical for user trust and engagement

‍

The Solution: AI-Powered Tags

To tackle these challenges, we collaborated with product, development, and AI teams over six months to design and implement AI-powered tagging. This feature uses AI language models to automatically generate, organize, and refine qualitative data labels, significantly improving research efficiency.

MVP Features

• Generate AI-powered qualitative and quantitative summaries.
• View, edit, and refine AI-generated tags and summaries.
• Share summaries and provide feedback for continuous AI improvement.

‍

V1 Enhancements

• Regenerate and manage multiple summaries for comparison.
• Filter and rename AI-generated summaries for better categorization.

‍

V2 Enhancements

• Customizable summary types tailored to specific research needs.
• Expanded tagging and theme management for enhanced research depth.

‍

User Research & Design Process

To ensure usability and effectiveness, we conducted extensive user testing, iterating on the design through multiple feedback cycles. Key design considerations included:

• Prioritizing tags and themes as primary elements in the response view.
• Establishing a hierarchy where objectives define AI scope, with organized tags and themes.
• Transparency in AI decisions, ensuring users understand how tags are generated and providing citation tracking for AI-generated summaries.

‍

User Research for use of product:

‍

‍

Prototyping Usability Testing for Various user flows:

‍

Conducted many user tests to iterate on current designs:

‍

Results & Impact

Fuel Cycle officially launched AI-powered tags on February 4, 2025, transforming how qualitative research is conducted.

• 10x Efficiency Gains: AI-powered tagging reduced qualitative data processing time from weeks to minutes.
• Improved Adoption: High customer adoption with increased usage across research teams.
• Actionable Insights: AI-driven summaries provided clarity, allowing stakeholders to derive insights faster.
• Enhanced Customization: Researchers can fine-tune AI-generated tags for specific research needs.

‍

Customer Feedback

A research team utilizing the feature shared their perspective:

“AI-generated tags transformed our workflow, automating a previously time-consuming process while allowing us to tailor the results to our needs. The balance of speed and accuracy is a game-changer.”

‍

‍Conclusion

The AI-powered tags feature is a milestone in Fuel Cycle’s Research Engine, redefining qualitative research with autonomous insights. By leveraging AI, we’ve enabled businesses to extract richer, more actionable insights at scale, reshaping how companies understand their customers.This project exemplifies Fuel Cycle’s commitment to innovation, efficiency, and user-centric design—paving the way for the future of AI-driven research.

Read about our release HERE.

‍

‍

‍