As development teams improve AI coding, the bottleneck for most enterprise engineering teams at scale is shifting to operations.  Leading engineering teams are reshaping teams and tools around the advantages that AI represents.  Most teams are seeing incidents and operational burn-out on the rise, and lagging teams are not yet seeing any AI impact yet.  This difficult middle state - where coding gets dramatically more efficient but it is surrounded by legacy operational practices - has been a challenging one for teams across the industry.  This talk will discuss some of the techniques we are seeing leading enterprise teams use at scale to move through that difficult middle state quickly and emerge on the other side with some of the most efficient “software factories” in the industry.