By now, many of the tools and techniques for tuning and improving database performance-such as design, schema updating, software upgrades, query optimization, hardware upgrades, in-memory processing, indexing, load balancing, sharding, and caching-are well-known and in constant use. Lately, however, a new challenge has entered the arena-and it requires new ways to assure performance within data environments. The rise of large language models (LLMs) to support business decision making and customer communications means databases are being pressed into service in new and highly demanding ways. LLMs are available publicly through services such as GPT-4 or are increasingly being built within enterprise settings for security and contextual reasons. While companies are still in the early stages of Al adoption, LLMs loom in their future as pressure mounts to deliver Al across a range of business functions.