Ever wonder how software developers truly know if you love their product? Traditional star ratings and reviews only scratch the surface. This research dives into a fascinating new approach: using Large Language Models (LLMs), like the tech behind ChatGPT, to gauge implicit user sentiment—those unspoken feelings that reveal true software desirability. The study uses the Microsoft Product Desirability Toolkit (PDT), which asks users to pick five words describing their experience with a software product. Instead of relying on explicit ratings, researchers fed these word choices and user explanations into several LLMs, including GPT-4 and Claude, as well as traditional sentiment analysis tools. The LLMs were tasked with generating a sentiment score from 0 to 1, representing negative to positive sentiment, for each user’s word groupings. Surprisingly, the LLMs excelled at this nuanced task, outperforming traditional methods. They provided not only numerical scores but also confidence levels and explanations for their assessments, offering a deeper, more human-like understanding of user feedback. This approach opens doors to understanding what users truly desire, even when they can't articulate it perfectly. Imagine a future where software anticipates your needs based not just on what you say, but on how you *feel* about using it. This research offers a glimpse into that exciting future, while also highlighting some intriguing observations. The study notes the importance of data formatting and order when interacting with LLMs, hinting at how these models process information. It also explores the usefulness of LLMs providing confidence levels in their judgments, giving developers a clearer picture of how much they can trust the AI’s evaluation. While more research is needed, this study lays the groundwork for a universal tool that can quantify those all-important implicit feelings, leading to the development of even more user-friendly software.