Data segregation requirements fundamentally break the utility of staged environments. There are some startups that try to fill this gap (Tonic.ai etc.) with data generation technologies -- yet it is extremely expensive to generate meaningful test data to populate development and staging environments that deeply mimic the interrelationships inherent to production data. These data relationships are very valuable to test. Because of this segregation, many organization resort to adhoc and manual testing techniques.