When you add a file, a page, or a note to a collection, Raison processes it so an assistant can look it up later. This page follows that path. It is not a legal inventory, and it does not replace the article on model training.
Where the content lives
Add resources in the library, or inside a collection under knowledge bases: files from your computer, a web page, pasted text, an LMS course, a SCORM package, or a hosted video. Those items become resources in a collection.
A resource in the library does not, by itself, change what an assistant answers. Link the collection on the assistant knowledge base.
What processing does
Raison reads the resource and builds a searchable version: extracted text, plus an index the assistant can query. Wait until the status is Ready. While it is processing, waiting, to review, or in error, that file is not reliable search material.
In chat, the assistant looks up the most relevant excerpts from the attached collections. It does not paste every document into a reply. How that search shapes the answer is in How Raison grounds answers in your course materials.
What learners can see of your materials
Learners do not browse your library. When an answer cites course materials, they can open Sources only if resource sharing is on (it is on by default). That control lives in learner experience settings.
A file a learner attaches in a chat is used in that conversation. It is not added to the course knowledge base. Notes in the notebook stay the learner's unless they share them with this assistant. Where notes and the knowledge base disagree, the knowledge base wins.
Optional extra sources
Course collections stay first. Two settings can widen where the assistant looks after that, if the organization allowed them and you turned them on: general knowledge, and web search on a given message.
Training
Raison uses your resources to answer in this product. It does not use them to train base language models. See Is my work used to train models?.
Related
Need help? Write to [email protected].