Explanation:

You are using the Azure OpenAI .NET SDK to send a prompt ("What is Microsoft Azure?") to a text model deployment. For text generation, you must call the Completions endpoint; hence client.GetCompletions(...) (or the async variant) is correct.
The Response < Completions > object returns one or more choices; the generated text is contained in response.Value.Choices[0].Text , which you can write to the console.
GetEmbeddings would return vector embeddings, not natural-language text, and GetImageGenerations is for DALL E image generation, so neither meets the requirement to "answer user questions" and print text.
Microsoft References
Azure OpenAI Service (.NET) - Completions quickstart and sample showing GetCompletions and reading Choices[0].Text .
Azure OpenAI Service - Embeddings overview (returns vectors, not text).
Azure OpenAI Service - Image generation overview (DALL E, not for text answers).