Clickthrough on search results have been successfully used to infer user interest and preferences, but are often noisy and potentially ambiguous. The reason mainly lies in that the clickthrough features are inherently a representation of the majority of user intents, rather than the information needs of the individual users for a given query instance. In this paper, we explore how to recover personalized search intent for each search instance, using a more sensitive and rich client-side instrumentation (including mouse movements) to provide additional insights into the intent behind each query instance. We report preliminary results of learning to infer query intent over rich instrumentation of search result pages. In particular, we explore whether we can automatically distinguish the different query classes such as navigational vs. informational queries. Our preliminary results confirm our intuition that client-side instrumentation is superior for personalized user intent inference, and suggest interesting avenues for future exploration.