Using a large language model is like talking to a personal assistant, confidant, topic expert, and copyeditor all at once: and that's a problem. By optimizing over many possible users, situations, and objectives, today's pre-defined AI systems ultimately produce generic AI interactions. In this talk, I argue that just-in-time architectures can produce specialized AI interactions that enable users to redirect generic AI systems to realize their specific, distinctive goals. I demonstrate this concept by introducing just-in-time (JIT) objectives, which specialize outputs to a particular user by inducing user objectives from observed interaction traces. JIT objectives enable on-the-fly generation of software tools that visualize a research statement's logical argument, test alternate color palettes for a figure, or provide feedback based on relevant academic experts. Then, I demonstrate how this approach also enables interventions into other AI systems such as social media rankers (Societal Objective Functions) and concept induction from unstructured text data (LLooM). This work argues that just-in-time AI interactions are a viable strategy to expand user control and combat the issues of generic AI interactions.