12 Voice AI Monitor

Design rationale: The card feed is the most natural mobile pattern for reviewing a list of items with rich detail. Each card is self-contained: quality score circle provides instant visual grading (green/orange/red), the transcript preview shows the first two lines for quick assessment, and actions are always visible. The escalated card has an orange border to draw attention. The 7-day sparkline trend in the header gives the manager a sense of whether AI quality is improving or degrading without needing to open a separate analytics view. Trade-off: only 2-3 cards visible at once on a phone screen.

Design rationale: The compact list shows all 8 calls at once, giving the office manager an immediate overview of the day's AI performance. The quality badge circle next to each row provides visual scoring. The expand-in-place pattern keeps context — the manager can read 4 lines of transcript without navigating to a new screen, which is enough to assess whether the AI handled the call well. The orange/red left-border on escalated and issue calls draws the eye to problems. Trade-off: the expanded preview is constrained to a small space and can't show the full transcript — the manager must tap "Full Transcript" for deeper review.

Design rationale: The split view shows both the call list and the full transcript simultaneously — no navigation needed. The quality breakdown (Greeting 10/10, Clarity 9/10, etc.) gives granular scoring that helps the office manager understand exactly where the AI excels or struggles. The full transcript with color-coded AI lines (blue background) vs patient lines makes it easy to follow the conversation flow. Trade-off: on a 375px phone screen, this is dense. The call list is compressed to ~140px height, showing only 4-5 rows before scrolling. Might work better on a tablet or in landscape mode. The office manager who really needs to review transcripts may prefer rotating to landscape.