Ask Campaign Assistant with Source Context

Use Ask when you need an explanation, continuity check, source review or summary. Use Propose when you want the Assistant to prepare a typed campaign change for separate review. Propose does not bypass review and does not apply a change by itself.

The mounted depth controls show the current quote before submission:

• Focused: 1 Gold for the active note or immediate context.
• Campaign-wide: 3 Gold for a wider search across campaign records.
• Deep Prep: 5 Gold for the dedicated Session Prep brief workflow.

The retained production journey spent 22 Gold across eight settled runs. That is evidence for this walkthrough, not a permanent price promise. Check the mounted quote again for each new request.

The composer shows the active note or surface, the chosen scope and the number of available sources. If selected prose is included, it also shows Selection. Remove unrelated material from the note or narrow the request if the context is broader than you intended.

Campaign Assistant can label content with four evidence classes:

• KNOWN is supported by campaign records.
• INFERRED is reasoned from the cited records.
• PROPOSED is a creative option and is not canon.
• DISPUTED means the campaign records conflict or remain ambiguous.

Do not flatten these labels into one answer. A useful inference can guide your review, but it is still different from a known campaign fact.

Choose a source chip to open Used campaign sources. Each entry can show its source key, title, record type, excerpt and inclusion reason. The retained privacy-safe crop proves those mounted fields for its S1 campaign record. It does not show optional session or transcript metadata.

Wait for the run to reach its terminal state. A visible answer, a completed run and a settled Gold entry are separate checks. The retained eight runs completed and settled for their displayed quotes, with no extra charge for the later local review actions.