For functional materials that are in a nascent stage, such as high-speed spintronics, quantum information storage, and new semiconductors, it is not clear what will be the high-performance materials of tomorrow. I will explain how we use chemical rules and bonding networks to identify promising uninvestigated or undiscovered materials. Our work shows that chemical spaces in emerging technical areas hold a surprising number of novel compounds. Because these materials have complex electronic, optical, and magnetic properties, they have anisotropic properties and growing large single crystals can be a crucial step toward understanding their behavior. The immediate impact of this framework is identification of spin behavior in Fe2As and Mn2Au, and materials where quantum information can be stored in the excitations of Eu3+ rare earth ions. Even with promising materials in mind, the task of crystal growth is still daunting. Due to the required millimeter dimensions, they must be grown from solutions, fluxes, or vapors. This process is often hard to observe, and highly kinetically dependent, so in situ techniques can be especially valuable. With a clearer view of how materials form, we can critically evaluate computational predictions (ab initio or machine-learned methods) and explore novel reactions to target new phases.