Quality Assurance

The QA module lets supervisors, QA leads, and quality analysts evaluate agent interactions, calibrate scoring across the team, and track quality trends over time. Agents see their own results in a dedicated self-view.

Open it from the sidebar under Quality → QA. The QA section has five top-level tabs:

Tab	What's there
Dashboard	At-a-glance KPIs, daily activity, top agents, recent submissions, deep links into reports
Scorecards	The evaluation templates — sections, questions, scoring rules, lifecycle (draft → published → archived)
Evaluations	The list of evaluations across the tenant, with filters by status and LOB
Calibrations	Group-scoring sessions used to align evaluator standards
My evaluations	An agent's own evaluations — review scores, acknowledge, dispute

QA dashboard

Roles

QA features are gated by permission strings under the qa.* namespace:

Permission	Typical role	Lets the user…
`qa.scorecard.create` / `.update` / `.archive`	QA lead	Author and lifecycle scorecards
`qa.evaluation.create`	QA evaluator, supervisor	Start a new evaluation
`qa.evaluation.read`	QA team, manager	See all evaluations in scope
`qa.evaluation.read.own`	Agent	See only their own evaluations
`qa.evaluation.acknowledge`	Agent	Acknowledge a submitted evaluation
`qa.evaluation.dispute`	Agent	Dispute an evaluation within the dispute window
`qa.evaluation.resolve`	QA lead	Resolve a disputed evaluation
`qa.evaluation.void`	QA lead	Void an evaluation (terminal state)
`qa.calibration.manage`	QA lead	Create and run calibration sessions
`qa.calibration.participate`	QA team	Score in a calibration session
`qa.report.read`	QA team, manager	Open the QA reports

The QA tabs render for everyone, but actions and data scope are enforced server-side — agents who reach /qa/evaluations will see only what they're permitted to see.

Roles​

What to read next​

Roles

What to read next