Today I thought a bit about something that I noticed several times now, in different places and companies. There is some big, old, monolithic monster to be slain and replaced by something new and better. The reason usually is to reduce maintenance costs and to speed up feature delivery time. However, it takes several failed tries (equaling several multi-year-projects) to succeed (if they succeed at all). Each try, new weapons are used: Agile, Microservices, Streams, whatever. Sometimes, strategies are changed, e.g. slow death or a big bang.
Often, the only result is the gained insight that all used weapons are useless and do not fulfill the expectations. And this insight is indeed true. Just calling people pigs and chicken, or using bleeding edge v0.x-Software will not kill any monsters.
What is needed to succeed is to actually change the company's habits in some significant way. The biggest point is to trust in people. Let them be responsible and accountable for their area of work, and they will put the available weapons to best use, and succeed. As the agile manifest states, "people and interactions over processes and tools".
Insisting on complex processes will slow down everyone. Focusing on "people and interactions" (synonym to "getting work done") will improve development speed by at least an oder of magnitude. Of course, having lightweight processes designed to help people is also necessary. The important thing here is that processes should be designed focusing on helping people. As a side effect, they may fullfill regulatory requirements, instead of just being there because of these requirements.