Context is the assumption you give to the AI search engine/assistant to presume in answering your question/request, it may be in two forms:

• Natural language sentences. This will be used in search/generating sentence stream to simplify repetitive input in your request.
• A node in the private/public space which might contain one or more sub-nodes and some learned content as well. The search or generating work will be based on these sub-nodes and content. In this way, you could manage your data into organic, logical manner to facilitate both your development and AI assistance. This could also enable you to scale up data to extremely high level(billions or trillions). This is something you can hardly find in any other AI companies in the world. The knowledge is accumulated, refined, and abstracted into a higher dimension and such  recursion will be your strong suit as time goes by.