What counts as RAG?

I have always considered the term RAG to be a hype term. to me Retrieval Augmented Generation just means the model retrieves the data, interprets it based on what you requested and responds with the data in context, meaning any agentic system that has and uses a tool to read data from a source (weather it's a database or a filesystem) and interprets that data and returns a response is technically augmenting the data and generating a result, thus it is RAG. Mainly just trying to figure out how to communicate with those that seem to live on the hype cycle