Your Rubric. Your Thresholds.
Your Coaching Voice.

Rewrite the full rubric text used to grade Sales, Support, Cancellation, Tech Support, and Warm Transfer calls — or add company-specific criteria like PIN-only verification.

Assign each rubric criterion a numeric weight (e.g. "Empathy = 30, Clarity = 20") so the AI-computed overall score is a true percentage-weighted average instead of a simple mean — with a live percentage preview per criterion.

Define required or forbidden phrases, write plain-English AI grading instructions per rule, and mark a rule to be skipped on outbound calls.

Decide how serious a failure of each rule is — Critical, Major or Minor — instead of leaving it to the AI. A Critical failure caps the call score in the Critical Fail band, whatever the scoring calibration is set to.

Choose how hard the AI applies its score caps — Lenient, Balanced or Strict — with the evidence beside the setting: how often managers who corrected a score raised or lowered it, per rubric metric.

Pick the model used for transcription, analysis, scoring and coaching (higher-tier models on Compliance IQ and above), and opt in to a fallback model that takes over when the selected one is out of capacity. Off by default, so your chosen models are never swapped silently.

Tune the tone of five separate AI personas — scoring metrics, scoring rubric, Coaching Hub, live training, and immediate coaching feedback — so results sound like your QA team, not a generic template.

Set the overall score threshold, medium/strong score bands, the auto-coaching trigger score, an optional per-metric threshold, and which of 7 signals (low score, missed opportunities, warm transfer, churn risk, and more) can fire Immediate Coaching.

Create teams and assign each one a specific scoring rubric and dashboard type (Customer Service or Tech Support), so different departments are graded and monitored on their own terms.

Link each team to one or more phone-system queues, so the Realtime Dashboard can filter by team and service levels are tracked per queue rather than blended.

Pull new recordings from your phone system automatically every 30 minutes, hourly, or every 2 hours.

Set thresholds for wait time, calls in queue, abandon rate, service level, break and lunch overruns and more, per queue where needed — and the days and hours of operation outside which phone-system alerts stay quiet.

Set company-wide targets for Avg. Score, CSAT, Resolution Rate, Handle Time, Call Volume, Close Rate, and per-rubric-metric goals — each individually toggled on or off for the My Performance page.

Control how many consecutive days in the Critical burnout tier trigger the first alert, and how often follow-up alerts re-fire while the condition persists.

Save a named set of notification preferences at the company level and apply it to new hires during onboarding.

Control which KPI tiles appear on the Realtime Dashboard, the live-poll interval, alert de-duplication window, and default follow-up event timing.

Require specific call metadata (agent, team, direction, recording) before a call enters the AI pipeline, and set duration caps that route outliers to manual review.

Filter call records, coaching sessions, and more by any combination of fields, choose the columns to include, and export or email the results as CSV — including scheduled digests.

Build recurring coaching-email digests scoped to a team or a specific agent, filtered by score range, rubric type, call direction, and compliance status.

View and act on Coaching Hub, dashboards, and call queues filtered to their own team — configured by an admin's team assignment, not self-serve.

Get alerted when calls about one topic — or one topic in one area — jump well above their usual weekly level in their own team’s calls, instantly or in a daily or weekly digest.

Silence agent-status or queue alerts for themselves, or have them emailed too, without changing the alert rules the rest of the company works to.

Same personal controls as agents (below) — enable/disable, delivery channel, and frequency — for every manager-facing alert.

Turn each alert type on or off, choose in-app, email, or both, and set the delivery frequency to instant, daily digest, or weekly digest.

One switch routes every notification to in-app only, without losing the underlying per-alert configuration.

Preview and apply another teammate's notification settings, or an admin-published template, before saving.

Download personal notification settings as JSON and re-import them — useful when moving between companies or restoring a known-good configuration.

Owns rubrics, compliance rules, coaching thresholds, AI personality tuning, team-to-rubric assignment, and performance goals — the settings that define how the whole company is scored.

Builds custom reports and export schedules, sets up coaching digests for their team, and manages their own notification preferences.

Controls personal notification delivery and frequency only — scoring rubrics and coaching thresholds are set above them, consistently, for fairness across the team.

A read-only reviewer. Sees what a manager sees, but cannot change any setting or record or start any AI generation.

For network and product engineers: Customer Themes and its alerts, nothing else — counts, trends, areas and spikes, never customer quotes or call links.