The new GPT models are triggering a revolution in the IT industry for its capability to generate text and images conforming to the typical human expectations. Anyway, sometimes you may have experienced results from ChatGPT that are not totally satisfying. The main reason relies on the Prompts that you use to instruct the service. Prompt Engineering is a new emerging discipline that helps in establishing the best way to interact with generative AI models using natural language and obtain the best results. In this session we will see the basic principle of Prompting, leveraging both public and private data using the advanced capabilities of the Azure OpenAI service and the Semantic Kernel library to create the next generation applications. We will see practical examples of code and prompts to generate apps interacting with different tones, languages (translation), roles (restaurant virtual waiter), and limiting responses to custom private data.