Not every modernization effort can be the aspiration of the autonomous agent, and not every legacy impediment can be safely delegated to autonomous agency agents. The choice is made for those workflows wherein agentic AI can provide outsized leverage—threats large but patterned code bases, with risks containable by policy; and measurable outputs-cycle time, change-fail rate, SLO adherence, cost—that can be improved through the tool-assisted intervention ad infinitum. This chapter will act as the on-field guide in locating those workflows and preparing them for the oversight of the agent.