Findings
Findings are AI-extracted observations from your meetings — organised by type, attributed to their source, and accessible from every Execution Scope.
What it is
When the Transcriber captures a meeting, the Meeting Intelligence Pipeline analyses the transcript and extracts structured observations about what happened: decisions made, actions committed to, risks raised, questions left open. These observations are classified and surfaced as Findings — curated, typed records that accumulate across all your meetings and can be managed, filtered, and acted on over time. Unlike a meeting summary, which is a narrative of what occurred, Findings are persistent and tied to specific participants and moments. Each finding is traceable back to the specific meeting it came from — and to the participant who raised it.
How it works
In Parallel classifies each extracted observation into one of seven types. Within each type, findings move through a lifecycle of states — from initial capture through to resolution or archiving. The detail view for each finding shows the AI-generated description, a direct quote from the participant who raised it, the source meeting, and a Properties panel with the finding's current state, owner, and scope.
Every finding is linked to the Execution Scope it belongs to, so the right team sees the right signals — not a cross-scope flood.
The seven finding types
Each type captures a distinct category of execution signal. Each list has a status filter bar — single-select, defaulting to the first lifecycle state for that type. Click a different status to switch, or click All to see everything.
Type | What it captures | Lifecycle states |
Action Items | Tasks and commitments made in the meeting | Backlog → Assigned → In Progress → Done → Archived |
Decisions | Choices confirmed or directions set | Proposed → Reviewing → Approved / Rejected |
Opportunities | Ideas and potential improvements worth pursuing | Identified → Qualified → Prioritized → In Progress → Won → Archived |
Escalations | Issues requiring urgent attention or resolution | Raised → Acknowledged → Investigating → Mitigating → Resolved → Closed |
Open Questions | Unresolved questions awaiting answers | Asked → Assigned → Answering → Answered → Closed |
Learnings | Insights worth retaining for future reference | Captured → Curated → Published → Applied → Archived |
Obstacles | Blockers or risks that could delay progress | Active → Archived |
Each list has a status filter bar at the top. The example below shows Action Items with Backlog selected by default — the first lifecycle state for this type.
The detail view
Opening any finding shows the full detail: an AI-generated description, a direct quote from the participant who raised it, and the Properties panel on the right. The quote block includes attribution — who said it, which meeting, and when — so the finding is always traceable to its source conversation.
The Properties panel shows: Type, State, Owner, Team (the Execution Scope), and Source (the meeting and date). Action Items also show a Due date and a Delegate button — delegating moves the item to another person's Personal Dashboard task list while keeping you notified of progress.
How to access Findings
In the left nav, select your scope and click Findings. This opens a drill-down showing all seven types. Click any type to see its list. Use the ← back arrow in the nav panel to return to the scope view.
What to do with each type
Findings are only useful if they're acted on. A few patterns that work:
Action Items — Assign an owner and due date. The item will appear in the owner's Personal Dashboard task list.
Decisions — Advance the state as the decision progresses (Proposed → Reviewing → Approved). This creates a living decision log for the scope.
Escalations — Use the state workflow to show active management (Raised → Investigating → Resolved). Escalations with no state movement signal issues being ignored.
Open Questions — Assign them so they don't disappear. Revisit in your next meeting.
Learnings, Opportunities, Obstacles — Review in planning sessions. These become the institutional memory that prevents teams from repeating the same mistakes.
Related