AutoQA Evaluations enable managers and supervisors to automatically grade agent performance without manual review. The AI evaluates every recorded and transcribed call — and, with chat transcript processing, every chat — against defined criteria covering adherence to greetings, processes, upsell attempts, closures, or any other measurable behavior, and produces a score that is available in reporting.

Multiple templates can be created and scoped to specific teams, skills, call directions, or channels.

📘
Reviewing and Reporting on AutoQA Results
Results can be reviewed per interaction in C2G — Examining the Call and compiled across interactions in reports, including the AI QA standard reports such as Agent QA Summary.

Requirements

QA Evaluation licensing is required; it is included with Elite licensing
Recording and transcription must be enabled for all calls to be evaluated

Key Considerations

Evaluations are processed after a call concludes and are generally available within a minute
Changes to an evaluation template apply to future calls only — scores for previously evaluated calls are not retroactively updated
Evaluations are based on transcript content; the AI does not detect audible tone directly, but can infer tone from dialogue
Timing-based criteria are not supported — the AI cannot detect whether a specific action occurred within a defined number of seconds, but can assess whether it occurred early or late in the call
Evaluation results support manager feedback (positive or negative) to fine-tune AI scoring over time
Each agent on a call is evaluated separately — a call handled by two agents produces one evaluation per agent, and each result names the agent who was graded
Voicemail recordings, for both direct and skill voicemail, are not evaluated

Enabling an Agent for Evaluation

Before a template can produce results for an agent, transcription and QA evaluation must be enabled at the agent level.

Navigate to Agent Licensing or UC Phone Licensing from the User Management menu
Find the applicable agent and enable both Transcription Enabled and QA Evaluation Enabled
Save changes

Accessing QA Evaluation Templates

QA Evaluation Templates is accessible to accounts with Supervisor or Admin permissions.

Navigate to AI Configuration > QA Evaluation Templates
All configured evaluation templates are listed in the modal

The CX QA Campaign Template

Sites licensed for both transcription and QA Evaluation start with a ready-made template, CX QA Campaign, so evaluations do not have to be built from scratch. It arrives disabled — no interactions are scored until it is enabled — and it can be edited like any other template. If it is edited or deleted, it is not restored on later updates.

Evaluation Point	Answer Type	Points
`Opening`	Yes/No	5
`Listening Skills`	Scale	10
`Empathy`	Scale	10
`Single-Call Resolution`	Yes/No	10
`CSAT`	Scale	5

The template covers inbound and callback calls with a 30-second minimum recording length, and is not limited to any skill or role. The AI QA standard reports — Agent QA Summary, Agent QA Time Interval, and Agent QA Calls — report on this template.

Configuring an Evaluation Template

Select New Evaluation Template to create a new template, or select an existing template to edit
Configure the template criteria:

Field	Description
Template Name	The name of the evaluation as it appears in reports and criteria (e.g., "Support Playbook")
Minimum Recording Length	Excludes calls shorter than the specified duration from evaluation. Useful for filtering out calls with insufficient content for meaningful scoring
Call Direction (optional)	Limits evaluations to calls of specific directions, such as inbound, outbound, or callback. The default considers all directions
Media Channel (optional)	Limits evaluations to specific channels, such as voice calls or chats. The default considers all channels
Skill (optional)	Limits evaluations to calls involving the selected skill(s). The default considers all call types
Agent Role (optional)	Limits evaluations to agents assigned the selected role(s). The default considers all agents. See Manage Roles

Save changes

Adding Evaluation Points

Evaluation points define the individual behaviors or tasks the AI should assess for each call. Each point is scored independently.

Select New Evaluation Point
Configure the point:

Field	Description
Evaluation Point Name	A short label for the metric. Displayed as a column header in reports — keep it concise (e.g., "Proper Greeting", "Upsell Attempt")
Full Question and Description	A detailed description of what is being measured. More descriptive criteria produce more accurate AI scoring. Example: "Did the agent professionally greet the customer and state their name as well as the company name?"
Answer Type	How the AI grades this point (see below)
Max Possible Points / Points for Yes	Applies to Yes/No and Scale types only

Answer Types

Type	Behavior
Yes/No	Full points awarded for "Yes"; zero for "No"
Scale	AI awards a score between 0 and the configured maximum (1–10), based on how fully the criteria were met
Text Response	AI provides a written reply (e.g., "How could the agent be more professional?"). Text responses are not included in score calculations or KPI formulas

Apply changes

Additional evaluation points can be added to the same template by selecting New Evaluation Point again. Points can be reordered by dragging the drag icon next to the point name.

Managing Existing Templates

Editing a Template

Select the Edit icon inline with the template name. Changes apply to future calls only.

Enabling / Disabling a Template

The toggle next to the template name controls whether the template is actively producing evaluation results. A template must be enabled to generate scores for completed calls.

Archiving a Template

Archiving disables the template and removes it from the primary list without deleting it. Archived templates can be restored by selecting Archived Templates and using the Restore Template icon.

Deleting a Template

A template must be archived before it can be deleted. Once in the archived list, select the delete option to permanently remove the template. This action cannot be undone.