Walk me through the Reason + Act cycle. Why is it better for complex, multi-step tasks than a single long prompt?

How does asking the model to "Respond as a panel of three experts (a coder, a security lead, and a PM)" differ from asking it to "Respond as a Senior Engineer"?

How would you set up an experiment to prove that a new prompt version is actually 10% better than the old one? What metrics would you track?

In an API inference call, what is the functional difference between the System, User, and Assistant roles?