Work with agent checklists

AI-generated summary

Agent Copilot can display up to 10 checklists of seven items each during an interaction. You can use checklists to guide agent behavior, ensure compliance, and clarify requirements. Checklists are automatically evaluated on agent turns and Agent Copilot checks off items when the agent speaks or types text that matches the item. You can configure checklists to trigger either at the start of an interaction, or by specific intents.

PrerequisitesClick to expand

Genesys Agent Copilot can display up to 10 checklists for agents during each interaction. Each checklist can contain up to seven items. Administrators can configure checklists that guide agent behavior, ensure compliance, and clarify requirements. Agent Copilot automatically evaluates checklists on agent turns and checks off checklist items when the agent speaks or types text that matches the item. To perform the automatic evaluation, Agent Copilot uses the help text, which the administrator specifies under each checklist item. It does not use checklist item names for evaluation. You can configure checklists to trigger either at the start of an interaction, or by specific intents.

Notes:

If the number of applicable checklists for an interaction exceeds 10, Agent Copilot displays the first 10 checklists.
When more than one checklist is displayed for an agent, all checklists except the most recent one are collapsed. When Agent Copilot checks off an item from a collapsed checklist, a visual indicator appears in the checklist title.
To ensure compliance, you can configure Agent Copilot to check off checklist items only when the agent utters a specific phrase.

Best practicesClick to expand

Make sure that the help text language matches the checklist language. This improves performance.
Keep the prompts simple to help deliver more consistent results.
Clearly define the format, structure, and level of detail you expect in the responses.
Align each item directly with a measurable behavior or goal, for example, instead of “Be polite” use “Greet the customer by name and thank them at the end of the call”.
Organize checklist items by a theme or logic, such as Customer Interaction, Compliance, and Product Knowledge. This helps reviewers navigate and complete evaluations faster.
Limit checklist length. A concise list of a few meaningful items is easier to maintain and yields more consistent scoring.
Write checklist items in plain, action-oriented language. Use verbs such as verify, confirm, ensure, or evaluate instead of vague terms such as understand or consider.
Revisit your checklist regularly to ensure that it reflects current policies, product updates, or business priorities. Set a quarterly review reminder for quality managers.
Combine automation and human review. Use AI for factual, pattern-based checks and a human review for qualitative checks, such as those for empathy or tone.
Test your prompts and adjust them to fit your specific needs and workflows.
Run a pilot of the checklist before you use it in production. Test it with a few agents or evaluators. Gather feedback on clarity, relevance, and usability, then adjust your checklist.
Ensure that all reviewers interpret and evaluate the checklist the same way. To achieve this, you can provide trainings, run calibration sessions, or share scoring examples.

Sample promptsClick to expand

Customer empathy and communication

Language: English GB
Description: Ensures that agents demonstrate empathy, active listening, and clarity during customer interactions.

Item name	Help text	Automated checking	Mark as important
Greeting and Introduction	Verify that the agent greeted the customer politely and introduced themselves at the start of the interaction.	No automation	Yes
Acknowledges Customer Concern	Check if the agent recognized and validated the customer’s concern before providing a solution.	AI checking	Yes
Clear Communication	Ensure that the agent used clear, concise language without jargon or filler.	No automation	No
Positive Closing	Confirm that the agent closed the interaction with a friendly and professional tone, thanking the customer.	AI checking	No

Compliance and policy adherence

Language: English GB
Description: Helps supervisors confirm that agents follow internal guidelines and maintain compliance across customer interactions.

Item name	Help text	Automated checking	Mark as important
Identity Verification	Check whether the agent followed all the required steps to verify the customer’s identity before sharing information.	AI checking	Yes
Accurate Policy Information	Confirm that all information shared aligns with the latest company policy and terms.	AI checking	Yes
Sensitive Data Handling	Ensure that no personal or confidential information was exposed in the conversation.	AI checking	Yes
Escalation Procedure	Verify that the agent followed the correct escalation process when unable to resolve the issue.	No automation	No

Resolution and follow-up

Language: English GB
Description: Evaluates how effectively agents use product knowledge to diagnose issues and deliver accurate resolutions.

Item name	Help text	Automated checking	Mark as important
Confirms Issue Resolution	Check if the agent verified that the issue was resolved before ending the conversation.	AI checking	Yes
Documents Resolution Clearly	Ensure that the agent recorded the steps taken and the outcome in the system notes.	No automation	No
Closes Interaction Positively	Verify that the agent thanked the customer and confirmed if they needed further assistance.	AI checking	No

Click Menu > Conversation Intelligence > Quality Management > Agent Checklists.
To add a new agent checklist, click Create Checklist. The Checklist Editor window appears.
To edit an existing agent checklist, click its hyperlink in the table. The Checklist Editor window appears.
In the Checklist Name box, specify the name of the agent checklist.
In the Language box, select the language of the agent checklist.
Under Checklist items, click Add item.
1. In the Item name box, specify the name of the item.
2. In the Help text box, provide detailed instructions for the checklist item.
  If you want to enable AI Checking, help text is required.
3. Under Automated checking, select No Automation or AI Checking.
4. Under Mark as important, toggle Yes or No.
  If you toggle to Yes, the item appears in bold for the agent.
5. Click Done.
To add another item, click Add item again.
To go back and edit an existing item, click Edit .
To delete an item that is currently expanded, click Delete. If the item is collapsed, click Trash .
To reorder items, click Grip and drag an item.
After you add all items to the agent checklist, test your checklist in the right panel.
1. Paste a transcript in the Transcript text box.
  Make sure that the transcript is in the format suggested on the preview panel.
  Note: The transcript must not exceed 64,000 characters.
2. To test the checklist for the first time or after you make an edit, click Rerun .
  Agent Copilot checks off every item that it can detect in the transcript. It also displays the detected checklist item under each relevant utterance.
  Note: You can only run test the same transcript against the same unedited checklist three times.
After you test the agent checklist, click Save Checklist.
To surface the agent checklist during interactions, add a rule to an Agent Copilot with the Checklist action. For more information, see Configure Genesys Agent Copilot rules.

[NEXT] Was this article helpful?

Get user feedback about articles.

Item name

Help text

Automated checking

Mark as important

Greeting and Introduction

Verify that the agent greeted the customer politely and introduced themselves at the start of the interaction.

No automation

Yes

Acknowledges Customer Concern

Check if the agent recognized and validated the customer’s concern before providing a solution.

AI checking

Yes

Clear Communication

Ensure that the agent used clear, concise language without jargon or filler.

No automation

Positive Closing

Confirm that the agent closed the interaction with a friendly and professional tone, thanking the customer.

AI checking

Item name

Help text

Automated checking

Mark as important

Identity Verification

Check whether the agent followed all the required steps to verify the customer’s identity before sharing information.

AI checking

Yes

Accurate Policy Information

Confirm that all information shared aligns with the latest company policy and terms.

AI checking

Yes

Sensitive Data Handling

Ensure that no personal or confidential information was exposed in the conversation.

AI checking

Yes

Escalation Procedure

Verify that the agent followed the correct escalation process when unable to resolve the issue.

No automation

Item name

Help text

Automated checking

Mark as important

Confirms Issue Resolution

Check if the agent verified that the issue was resolved before ending the conversation.

AI checking

Yes

Documents Resolution Clearly

Ensure that the agent recorded the steps taken and the outcome in the system notes.

No automation

Closes Interaction Positively

Verify that the agent thanked the customer and confirmed if they needed further assistance.

AI checking

Work with agent checklists

[NEXT] Was this article helpful?

Loading...

Work with agent checklists

[NEXT] Was this article helpful?