We are starting new AI project, to serve knowledge based on thousands of our company documents. We will be using Azure OpenAI,as the AI model and Azure AI Search as the vector DB. We tried some POC, spring AI in Java keeps things simple and easy to understand, since it has wrapper over Azure OpenAI as well. Microsoft has Semantic kernel and tons of features, to write AI applications, but very complicated to follow and understand. Is it safe to start our application with Spring AI, can we get all new features of Azure OpenAI and Azure AI Search using Spring AI or if not, using direct calls to Azure OpenAI. Or, is better to use Microsoft Semantic Kernel? Kindly share your experiences.