Score Every Agent
on Only Their Part of the Call.
Warm transfers, conference calls, and multi-agent interactions all involve more than one voice. Call Coach IQ separates audio at the waveform level before transcription, creates a scored record per speaker, and resolves warm-transfer participants to real user accounts — so every agent owns exactly their portion, nothing more.
Without Multi Agent Detection
- ✗The entire transferred call is scored as a single block — mixing IVR hold time, the first agent's segment, and the transferred-in agent's segment into one score.
- ✗The transferred-in agent's performance is diluted by — or credited to — work they had nothing to do with.
- ✗Speaker labels in the transcript are anonymous (Speaker 1, Speaker 2). Managers have no idea which segment belongs to which agent without listening back.
- ✗When auto-resolution fails, the record sits unassigned and scoring stalls.
With Call Coach IQ
- ✓Separate child call records are created per agent. Each record contains only that agent's audio segment, transcript, and score.
- ✓Each agent is scored exclusively on their own segment. Their scorecard, trend line, and coaching history reflect only what they said and did.
- ✓Warm-transfer speaker labels are automatically resolved to real user accounts. Managers see names, not speaker numbers.
- ✓When auto-resolution needs review, the manager is notified and an "Assign Agent" UI surfaces directly in the call record — resolution takes seconds.
How It Works
Acoustic Diarization Before Transcription
Audio is speaker-separated at the waveform level before any transcription occurs. Each speaker's audio is isolated into its own segment, ensuring the transcript and downstream scoring always know which voice said what.
Warm-Transfer Agent Name Resolution
Speaker segments are matched against real user accounts in your organization. Warm-transfer participants are identified by name, role, and team — not left as anonymous speaker labels.
Per-Agent Scoring via Child Call Records
Each resolved speaker gets their own child call record linked to the parent call. QA rubrics score only that agent's segment — producing a score that is fair, accurate, and attributable.
Agent-Visible Scorecards on the Parent Call
Agents can see their own scorecard directly on the parent call review page. They see their segment score, rubric breakdown, and any coaching or dispute filed against their portion — without seeing scores for other speakers.
Manager Notification + Manual Assignment UI
When automatic resolution cannot confidently match a speaker to a user account, the call's manager is notified. A one-click "Assign Agent" control in the call record lets them resolve the attribution without leaving the review.
Per-Segment Coaching, Dispute & Compliance Attribution
Coaching entries, agent disputes, and compliance findings are all tied to the specific speaker segment — not the call as a whole. Compliance teams can flag a violation against exactly the right agent, and agents can dispute only their own portion.
See Multispeaker Transcription in Action
Who Benefits from Multi Agent Detection
For the Customer
When your call gets warm-transferred, each agent you spoke with is scored and coached only on the part of the conversation they actually handled — so the person who resolved your issue gets recognized for it.
For the Agent
Your scorecard reflects exactly your segment of a warm-transferred call — never diluted by, or credited for, a colleague's portion of the conversation.
For the Business
Per-agent child records and automatic warm-transfer name resolution eliminate score contamination from transfers, so scorecards and coaching data stay accurate across every team.
Fair Scores for Every Agent on Every Call
Upload a real call and see per-agent scoring, warm-transfer attribution, and AI coaching in action — no account needed.
No account needed. Results in minutes.