> ## Documentation Index > Fetch the complete documentation index at: https://docs.riktra.com/llms.txt > Use this file to discover all available pages before exploring further. # Meetings > What happens to a transcript you upload. Paste or upload a meeting transcript and Riktra pulls out **candidates** — decisions, open questions, proposals, insights. You review them before anything becomes company truth. Nothing extracted from a meeting is written directly. Ingestion produces proposals like any other path into shared context; see [Governed truth](/concepts/governed-truth). ## What is deliberately withheld Sensitive material — health, pay, judgments about individuals, credentials — is never proposed as company context. It is withheld entirely rather than flagged for review, because a reviewer working through a list is exactly the wrong place to first encounter a colleague's salary discussion. This pass is one of the few places where a bug would be genuinely expensive, and it has been one. Swedish inflection defeated a word-boundary pattern during the build, which would have published pay discussion into company context. It is now covered by tests in both languages. ## What happens to the transcript It is kept only until its candidates have been reviewed, then deleted. Riktra is not an archive of what was said; it is a record of what was settled. ## Writing a prompt that extracts well If an agent is doing the extraction rather than the built-in pipeline, the distinction worth encoding is decided versus discussed: ``` Here is the transcript of today's meeting. Read it and propose what was actually settled as decisions. If something was discussed but not decided, propose it as an open question instead — never as a decision. Anything personal — health, pay, judgments about individuals — leave out entirely. ``` "We should maybe use Jira" is not a decision, and an agent left to its own judgment will often record it as one.