Decisions made in conversation usually disappear. Decision Capture watches your discussions and comments for decision signals, and when it finds them, it saves a decision record automatically.
How detection works
When a comment or discussion is created, Malve scores its text against weighted groups of decision signals: decision verbs, outcome language, and consensus phrasing. When the combined score crosses a confidence threshold, Malve creates a decision record and links it to the conversation it came from.
To be precise about what this is: it is phrase-based signal detection, not AI or natural-language understanding. It matches the shape of how teams write decisions ("we decided to", "let us go with", "agreed, we will"). That keeps it predictable and explainable, and it means you can rely on it without wondering what a model inferred.
Capturing manually
Detection is a safety net, not the only path. You can mark any discussion as a decision yourself when you want to record something the detector would not catch.
What you get
Every captured decision is a searchable record tied to its source conversation, so six months later you can find what was decided and the discussion that led to it. From there you can export it to Notion or Confluence.