Which platform would you choose for this data engineering scenario?

We're evaluating Databricks, Google Vertex AI, and Azure AI Foundry for building enterprise AI agents/chatbots over internal documents. On paper, all three seem pretty capable. I'm currently leaning towards Databricks because I like the idea of having the data, governance, vector search, and AI capabilities on one platform, but I'm not sure how much of that actually translates into a better experience in production. For those who've worked with two or more of these, which one did you end up choosing and why? Were there any capabilities (or limitations) that only became apparent once you were r