Using Large Language Models (LLMs) and a range of machine learning techniques—including clustering, dimensionality reduction, and sentiment analysis— AI Program fellows worked to organize unstructured customer feedback into 10–15 distinct, meaningful themes.
Published
04/16/2025
So, what do customers really think?
That’s the question Clorox challenged our AI Studio fellows to answer in their latest Challenge Project. With thousands of online product reviews at their fingertips, Break Through Tech fellows set out to build an advanced topic modeling framework that could cut through the noise and surface what matters most to consumers.
Using Large Language Models (LLMs) and a range of machine learning techniques—including clustering, dimensionality reduction, and sentiment analysis—fellows worked to organize unstructured feedback into 10–15 distinct, meaningful themes. Their solution not only identified key topics and the top keywords associated with them, but also pulled out representative reviews, analyzed sentiment, and mapped how Clorox products ranked on those themes compared to competitors.
The project was more than a technical challenge—it had real business implications. For Clorox, the ability to better understand customer feedback means more informed decisions across product development, marketing, and customer satisfaction. By refining and improving topic modeling capabilities, the students helped create a framework to track emerging trends, highlight pain points, and guide strategic investments.
To get there, fellows applied natural language processing (NLP) techniques and machine learning workflows to analyze and synthesize large volumes of consumer text data. They compared different modeling approaches and fine-tuned parameters to improve the clarity and relevance of the resulting insights. Deliverables included clearly defined themes, supporting review excerpts, sentiment trends, and comparative analysis against key competitors.
This Challenge Project is just one example of how Break Through Tech AI Program fellows bring immediate value to real-world business settings. With hands-on experience in AI and machine learning, Break Through Tech fellows are ready to contribute from day one—especially for organizations looking to translate unstructured data into actionable strategy.
Break Through Tech fellows have worked on more than 275 projects like this for companies across industries. If you’re looking to tap into the next generation of AI talent to tackle your own business challenge, contact us to learn more.
We’d love to work with you!