Imagine a search engine that not only understands your queries but also considers your personal background, like your interests or profession. This is the promise of personalized conversational information retrieval (CIR), and new research explores how to make this vision a reality. Traditional search engines consider the words in your query and return relevant results. CIR, however, aims to factor in the history of your search (the conversational aspect) and your unique profile. This could revolutionize how we find information, tailoring results to our individual needs. The key challenge? Personal textual knowledge bases (PTKBs), collections of text describing a user's background, can be quite noisy. Not every piece of information in your PTKB is relevant to every search. This research delves into different strategies for selecting the *right* knowledge from a user's PTKB and using large language models (LLMs) to reformulate search queries. The results? Interestingly, simply using all the information from a PTKB doesn’t always improve search results. However, LLMs excel at creating more tailored queries when provided with high-quality guidance on which parts of the PTKB are actually useful. The study suggests that LLMs can potentially crack the code of personalized search by intelligently incorporating the right bits of personal context. While the idea of personalized search has been around for a while, this research offers valuable insights into how we can move closer to truly personalized search experiences. Future research may involve even more sophisticated user modeling, going beyond individual sentences to create a comprehensive picture of user preferences. The road to truly personalized search is exciting, and these findings pave the way for a more tailored and intuitive way to navigate the ocean of digital information.