Most systems in production already have key use cases that stand in for unit and system testing. Find those and use them after every new feature or change to th

Most systems in production already have key use cases that stand in for unit and system testing. Find those and use them after every new feature or change to the system. I support an existing reporting system with several hundred tables and thousands of interdependent stored procedures. This system has a fairly often-used "global report" which touches nearly 90% of the tables. We run this user report after all changes to the database with a few swaps in parameters (history vs current etc). If this report and a few others "pass" (results are the same as previous reports), we know that the changes are safe. When the reports show differences, we confirm that the changes were intended. If not, something broke.