Starter
Quality Assurance
Build scorecards once and reuse them across clients. Evaluation lifecycle (draft → submit → dispute → resolve), bulk evaluation, calibration sessions with blind scoring, and a direct hand-off into the coaching module. Quality analytics roll up by LOB (Line of Business: a distinct client program or queue within an operation.), client, agent, and team. Sub-spec 30B (AI Quality Scoring) layers on automated scoring once the human baseline is in place.
What's inside
QA dashboard
Single landing page for QA leaders — submitted evaluations, average score, pass / auto-fail rates, daily activity, and weakest sections all in one place. Filter by client, LOB, and time period; jump straight to the underlying reports.
Agent QA portal
Each agent has a personal QA portal — every evaluation against their interactions, with score, pass / fail outcome, dispute window, and acknowledgement state. Submitted evaluations only; drafts stay invisible until the evaluator publishes them.
Scorecard builder (sections + questions)
Author scorecards with weighted sections, mandatory items, and conditional questions. Reuse a base scorecard across clients with LOB-specific tweaks; no more "which version of the scorecard are we using?"
Evaluation workflow (draft → submit → dispute → resolve)
QA analysts evaluate calls in a structured workflow with full audit. Agents can dispute; supervisors resolve; everything is timestamped. Calibration debates turn into structured conversations, not spreadsheet fights.
Quality analytics + reporting
Trend analytics by agent, team, LOB, client, and scorecard section. Calibration drift, evaluator-bias detection, and the QA scores that feed the SLA scorecard for clients.
Calibration sessions + blind scoring
Calibration sessions with multiple evaluators scoring the same call blind, then revealing and discussing variance. The platform records the variance and the resolution — calibration becomes data, not folklore.
QA → coaching recommendation linkage
A low score on a specific scorecard section automatically suggests a coaching session on that topic; the coaching module picks it up and the supervisor sees it in their queue. QA findings stop dying in a spreadsheet.
Bulk evaluation submission
CSV and JSON bulk imports for high-volume evaluation pipelines — auto-coaching tools, third-party speech-analytics platforms, and historical backfills land directly in the evaluation queue without manual entry. Async job processing per spec 30 §5.3.
AI Quality Scoring
AI-driven scoring of calls or interactions against your existing scorecard, surfacing top-priority evaluations for human review. Doesn't replace QA analysts; it triages so they spend their time where it matters most.