The author gives some interesting business-related examples, like invoice processing and covers many topics in prompt engineering, agents, etc. There are quite a lot of repetition in her examples -e.g., for every example she describes how to build almost the same UX. But the main issue, imho, is her avoidance of “hallucinations”. Except mentioning it a bit in prompt engineering chapter all her agents plan everything write, LLMs always provide a right result.