I think the Cynefrin frameworks obscures the most important distinction, which is between linear and complex systems. Complex systems can be simple or complicated, just as linear systems can be. The most important distinction is that a single intervention in a complex system does not necessarily produce any result, whereas a single intervention in a linear system (no matter how complicated) must always produce some result. That, and the recognition that almost all systems important for development are complex rather than linear.
I agree that we need eyes and ears on the ground to deal successfully with complex systems. That is why I am frustrated that the World Bank continues to rely primarily on super-educated technical specialists who occasionally drop in with "best practices" solutions, while it does not seem able to recruit and retain people who are willing to stick it out for the long term in hardship posts. This is rational from the private perspective of the individual experts, but not conducive to solving problems in complex systems.