Glossary

Every metric and control the platform reports — in the same words the screens use.

Structure

Workspace The container that owns calls, users, goals, QA rules and saved charts. Selected in the top bar; nothing crosses between workspaces. See Workspaces.
User Someone who signs in — an admin or a user. Users are assigned to specific workspaces.
Agent A person whose calls are audited. Agents live on their own roster and most never sign in; those who do see only their own calls or performance.
Org Agent ID Your own internal identifier for an agent, stored alongside their name so the roster maps back to your WFM or CRM.
Role Admin, user or agent — the top-level determinant of what someone can reach.
Agent access level No Access, Calls Access, Performance Access or Full Access — what an agent can see of their own data.

Call ID The platform's unique identifier for a call. Copyable from the call row; the thing to quote to support.
Connected source Storage the platform reads from — Amazon S3, Google Drive, Dropbox, Box, GCP Bucket or Wasabi.
Scheduled scan A source-plus-folder configuration that ingests new recordings every 4, 8, 12 or 24 hours, or on demand.
Run Now Triggers an immediate scan of a configured folder rather than waiting for the schedule.
Last Scan When a folder was last checked. Not Scanned means it has never run.
Processing status Queued (waiting), Running (in progress), Succeeded (available in the dashboard) or Failed (with a reason, and re-runnable).
Processing limits Per-workspace rules that exclude calls from processing: minimum duration, maximum duration, and the percentage of each agent's calls to process.
Metadata Your own fields attached to calls — campaign, queue, disposition, customer reference. Uploaded in bulk or edited per call, and filterable once attached.
PII masking Per-workspace redaction of personally identifiable details from transcripts. Applies to calls processed after it's enabled.
Primary language The transcription language for a workspace — English, or Other for non-English processes including Hindi and code-switched Hinglish.

Text sentiment Positive, neutral or negative, derived from what was said. Reported for the whole call and separately for agent and customer.
Overall sentiment The call's single sentiment verdict — the dial in the call list and the headline stat on the call screen.
Emotion Six classes derived from the audio: happy, sad, neutral, fear, disgust, surprise. Reported for the whole call and per speaker.
Why both Sentiment reads words; emotion reads tone. Polite words in a fearful or disgusted voice is the pattern manual sampling misses most.

Talk ratio / Percent Talk Share of the call the agent spent speaking.
Percent Silence Share of the call that was silence — the aggregate dead-air measure.
Dead air A silence longer than the workspace's threshold (configurable, 3–10 seconds).
Cross talk Both parties speaking at once — interruption. High cross talk on negative calls is a strong escalation signal.
Duration Call length in minutes. Shown in red on the call list when it exceeds the workspace's average-duration goal.
Call direction Inbound or Outbound, inferred from the recording's filename prefix where your recorder provides one.
Channel Agent side or customer side. Stereo recordings with speakers on separate channels give the cleanest separation.
Call summary A short written account of what happened on the call, with a listen-aloud option.

Flag A moment the platform marked as notable — risky, non-compliant or notably good. Flags are analyst-verified before they reach your reporting.
Severity grading How findings are ranked in the weekly report: Intolerable, Substantial, Moderate, Tolerable.

Call score / Overall Call Score The call's score out of 100, combining the three stage scores using the workspace's weightage. 90+ Excellent · 75–89 Good · 50–74 Average · under 50 Poor.
Opening score The call's opening, from Welcome, Name Verification and My Name.
Handling score The body of the call, from Courtesy, Cust Education, Active Listening, Assistance, Cushioning and Probing.
Closing score The close, from Interaction Closing.
Weightage How much each stage contributes to the overall score. Set per workspace, which makes scores comparable across your own agents but not across companies.
Customer Interaction Indicator The soft-skills checklist scored per call, with the reference words the platform listened for and a "n of m skills met" summary.
Reference words The phrases associated with a skill — what the platform listens for to judge whether the skill was demonstrated.
Parameter score On the performance screen: opening, handling and closing scored separately, so you can see which stage is weak.

Goals / targets Per-workspace targets — emotion mix, stage weightage, overall sentiment, calls per day, average duration, percent talk and cross talk.
Benchmark The comparison of a call or period against those targets — the strips on the emotion and talk panels, and the rings on the dashboard KPIs.
Goal column In the agent league table: calls achieved against calls-per-day multiplied by the days in range, with the percentage achieved.

Bookmark A per-user mark on a call for follow-up. The Bookmarks toggle filters the list to just those calls.
Saved filter A named advanced filter, reusable from the toolbar dropdown.
Shared filter A saved filter shared with colleagues, so a team reviews a consistent set of calls.
Comment & mention A note attached to a call, optionally tagging colleagues. Tagged people get a notification that opens that call.
Quadrant chart Two metrics plotted against each other, one dot per agent. The value is in the corners — typically high duration with low score.
Weekly report The PDF QA report delivered every Monday: executive summary, severity scorecard, critical calls with analyst notes, agent leaderboard, coaching and recognition, compliance trends, recommended actions, callback list and department risk summary.

Available minutes / Remaining duration Processing minutes left on the plan. Calls consume roughly their audio duration.
Available balance Funds on the account.
Ledger The day-by-day record of minutes and amounts credited and debited, expandable to individual entries.
Raise dispute Sends our billing team a message attached to a specific invoice.
Webhook URL Your endpoint for processing events, configured under Account → Preferences and validated with a Test before it can be saved.