Evaluate agent conversations for quality

When this is done, your supervisors or quality analysts can score agents' conversations against a quality script – during the conversation or afterwards. The agent receives the evaluation by email, and the score shows in Task HistoryTask History – A list of an Agent's past Tasks in the Agent Application, 7 days by default, filterable by media, direction and media-specific details such as email subject or phone number., so you can follow quality per agent over time.

Agent handles a call or chat
  → the evaluator scores it live from the monitor window,
     or later from Task History
  → the evaluator goes through the quality script to the end
  → the agent gets an email with a link to the evaluation; the evaluator is copied
  → the score and the evaluator's name show on the task in Task History

It takes about 20 minutes, once the quality script exists. Roles involved: Administrator (Part 1), Supervisor (Parts 2 and 4), Agent (Part 3).

Before you start

What	Who	Where
A quality script with [QM] in its name, built for your evaluation form	Graia – ask your Graia contact	–
Scripting turned on for your tenant	Graia	–
Call recording turned on for the channels you'll evaluate, so evaluators can listen to past calls	Administrator	–
A finished test call or chat from an agent, with completion code and notes	Agent	Your first shift in the Agent application

Only scripts with [QM] in their name (in any letter case) are offered for evaluations. Tell Graia which questions, scores and rating you want on the form.

Emails can be evaluated afterwards, but not live.

Part 1 – Give evaluators access (Administrator)

1. Let the role update historical reports

1. Go to Administration › Roles and open the role your evaluators have.
2. For Historical Reports & Insights, turn on Update as well as Read.
3. Click Save.

Expected result: the role has Read and Update on Historical Reports & Insights. Without Update, the evaluation button in Task History doesn't work.

See Edit a role.

2. Turn on scripting access for each evaluator

Evaluators need scripting access to run the quality script, even if they never handle tasks.

1. Go to Administration › Users and open the evaluator. If they're an agent or supervisor, go to Agents › Agents and open the User Details tab instead.
2. Under Scripting access, choose User.
3. Click Save.

Expected result: User is selected. If neither option is selected, the user has no scripting access and can't evaluate.

See Scripting access.

3. Make live evaluators supervisors

Only skip this if your evaluators only score past conversations.

1. Go to Agents › Agents and open the evaluator.
2. On Agent Configuration, set Type to Supervisor, and click Save.
3. Make sure they're in a team with the agents they'll evaluate. See Teams.

Expected result: the user is a supervisor. For other users the Monitor button in the Teams report is greyed out.

Ask evaluators to log out of the portal and back in so the changes apply.

Part 2 – Evaluate a conversation (Supervisor)

Score a past conversation, a live one, or both.

4. Evaluate a past call or chat

1. Go to Historical Reports › Tasks.
2. Turn on Show not evaluated tasks only, and use Add filter › Agent to find the agent's tasks.
3. On the task, click the evaluation icon (Click to start task evaluation). For a call, the recording starts playing at the bottom of the screen.
4. In Evaluate task, choose the quality script in Script used for evaluation and click Start evaluation.
5. Listen to the recording, or open the conversation history for a chat or email, and answer the script's questions.
6. Go through the script to its end.

Expected result: the evaluation is complete when you reach the end of the script. If Start evaluation stays greyed out, no script is chosen. If the list is empty, there's no [QM] script.

See Evaluations.

5. Evaluate a live call or chat

1. Go to Realtime Reports › Teams, open the team's Agents tab and click the agent's card.
2. On the open task, click Monitor. A monitor window opens; you listen in (call) or read along (chat) without the customer or agent knowing.
3. In the monitor window, under Live evaluation available., choose the script in Script used for evaluation and click Start Evaluation.
4. Answer the questions as the conversation goes on, and go through the script to its end.
5. Click Drop call (for a call) to leave, and close the window.

Expected result: you score the conversation while you listen. If the window says Evaluation is not available., there's no [QM] script, or you're missing access from Part 1.

See Monitor agent conversations and Watch your queues and teams live.

Part 3 – Read the evaluation (Agent)

6. Open the evaluation email

1. The agent opens the email that arrived when the evaluation was completed.
2. They click the link to open the evaluation.

Expected result: the agent sees Evaluation of <agent>'s task at <date> with the answers and the score, in read-only mode. The evaluator received a copy of the email.

Part 4 – Follow up (Supervisor)

7. Find the score in Task History

1. Go to Historical Reports › Tasks and turn off Show not evaluated tasks only.
2. Click Add filter › Evaluator and choose yourself.
3. To keep the results, click Export.

Expected result: the task shows the evaluation score and the name of the last evaluator. In the export, they're in the Total Score and Evaluated By columns. To evaluate the same task again, click the evaluation icon again – its tooltip now reads Task already evaluated.

See Export task history.

Check that it works

#	Do this	You should see
1	As the evaluator, open Historical Reports › Tasks and click the evaluation icon	Evaluate task, with your [QM] script in the list
2	Start the evaluation and go through the script to the end	The evaluation completes
3	Open the agent's mailbox	An email with a link to the evaluation; the evaluator is in CC
4	Filter Task History by Evaluator	The task, with the score and your name
5	As a supervisor, monitor a live call	Live evaluation available. in the monitor window

If it doesn't work

Symptom	Likely cause	Fix
The evaluation icon does nothing	The role has no Update on Historical Reports & Insights, or the user has no scripting access	Steps 1 and 2
Script used for evaluation is empty	No script has [QM] in its name, or scripting isn't on for the tenant	Ask Graia
Monitor is greyed out	The evaluator isn't a supervisor	Step 3
No Monitor button on the task	It's an email – emails can't be monitored	Evaluate it afterwards (step 4)
The monitor window says Evaluation is not available.	No [QM] script, or access is missing	Steps 1 and 2; ask Graia for the script
No recording to listen to	Recording isn't on for the channel, or it's a chat or email	Read the conversation history instead
Changes don't apply	The evaluator hasn't logged in again	Log out of the portal and back in

Related

• Task history: evaluations
• Monitor agent conversations
• Agents view of the Teams report
• Roles
• Users: scripting access
• Watch your queues and teams live, and step into a call