Licensing checklists

Knowledge base licensing: versions, access, and source checks

Prepare support articles for AI evaluation by separating tickets, public and internal content, revisions, product applicability, and source gaps.

A knowledge base can include public articles, internal procedures, and material copied from support conversations. Inventory those categories separately rather than proposing a complete system export. The goal here is to define a reviewable collection and a delivery plan before discussing commercial terms.

Separate articles from support conversations

A knowledge base export can contain public help articles, internal runbooks, draft notes, or copied ticket material. Inventory those categories before extracting a sample. Keep customer conversations, credentials, and account identifiers out of the exploratory brief. An article’s public visibility does not by itself resolve every permission or privacy question.

Record the content source and access category

For each article group, retain authorship or source history, supplier excerpts, and internal/public status. Link the records needed for permission review without embedding confidential contracts in a transferable package. Flag examples copied from customer cases and third-party documentation for separate consideration.

Resolve revisions and product applicability

Keep a stable article ID and a revision ID, plus product version, language, publication date, and deprecated status. Record replaced or conflicting procedures. Avoid flattening a revision history into several apparently current answers. Where applicability is unknown, label it unknown instead of extending instructions to all product versions.

Design retrieval tests around conflicting guidance

Use fictional or appropriately cleared example questions that distinguish current from obsolete procedures. Ask the evaluator to return the supporting article and revision, and to report when the archive lacks an answer. Include broken links and missing product coverage in the evaluation notes. These checks do not establish that a deployed support system will answer safely.

Describe updates and removals explicitly

State whether the proposed export is a snapshot or a recurring feed. Document how corrections, retired articles, and changed access categories are represented. Keep an exclusion list for private or unresolved material and verify it against the final export. An update plan should identify an owner and a feasible cadence before it becomes a commercial promise.

Knowledge base review sheet

Edit this outline, then copy or download it. It stays in this page and is not submitted or saved by this tool. Keep confidential records and personal information out. Save your work before leaving or reloading.

You can select and copy the text manually. This is a preparation worksheet, not a legal agreement.

Sources and scope of this guidance

Sources checked October 19, 2026. AWS describes field dictionaries as an aid to evaluating a dataset. The workflow and examples here are our preparation suggestions, not AWS requirements or a claim of platform eligibility.

AWS Data Exchange: data dictionaries