What belongs in a course Agent's Knowledge collection, and what should stay out?

The question

What criteria do you use to select documents for a course Knowledge collection, and which materials do you keep out?

Useful context

Knowledge collections store uploaded files, plain text snippets, web pages, and linked cloud-storage folders to ground Agent responses. In practice, faculty tend to converge on a similar split:

Commonly included Usually kept out
Course syllabi and lecture slides Unreleased exam drafts and answer keys
Open-access readings and rubrics Proprietary publications without reuse rights
Department policy documents Student education records

Management strategy is where practice diverges. Some instructors stage collections module by module so information unlocks as the term progresses, while others maintain one comprehensive collection and narrow the focus through Agent instructions or Knowledge Base Constraint, which offers prompt-level guidance rather than hard file access restriction. The choice usually comes down to how much of the term’s material you are comfortable having available in one place.

Share your approach

Describe how you manage course Knowledge collections:

  • Standard inclusions. Which document types go into your default course collection?
  • Deliberate exclusions. Which materials do you keep out, and why?
  • Update cadence. Do you revise weekly, or fix the document set for the whole term?
  • Retrieval quality. Which files does your Agent lean on most, and how do you handle a passage that comes back cited inaccurately?

Before you post

Before sharing prompts, configurations, or screenshots, review Community sharing and responsible AI use. Never post student records, personally identifiable information, or credentials.