About the Auto QM evaluation criteria

Auto QM uses AI to evaluate conversations. The percentage-based evaluation is based on set of questions that you select for your evaluation form. The evaluation scores are automatically generated and available at both individual and aggregated levels. When you build an Auto QM evaluation form, you can select questions from the following list:

Question category	Evaluation questions	Answer options
Accuracy & Transparency	(AI Agent only) Did the AI Agent self-correct itself when it made an error or provided inconsistent information? (AI Agent only) Did the AI Agent provide a response indicating that it does not know the answer to the customer’s question? (AI Agent only) Did the AI Agent provide an accurate answer to the customer’s question? (AI Agent only) Did the AI Agent appropriately communicate uncertainty when the customer is persistent?	Yes, No, N/A
Active listening	Did the agent successfully keep the conversation on track to focus on the primary issue and avoid diverging to other issues? Did the agent ask clarifying and relevant questions to better understand the customer's needs? Did the agent proactively guide the conversation, keeping it focused on the issue at hand? Did the agent summarize or paraphrase the customer's issue to ensure they correctly understood it? Did the agent actively commit to finding a solution and take ownership of the solutioning process for the customer's issue? (AI Agent only) Did the AI Agent repeat or paraphrase itself more than once in the conversation? (AI Agent only) Did the AI Agent provide a response that is relevant to the customer’s query? (AI Agent only) Did the AI Agent use earlier context from the conversation appropriately in later responses? (AI Agent only) Did the AI Agent adjust its style after detecting customer frustration or confusion?	Yes, No, N/A
Authentication	Did the agent authenticate the caller’s identity?	Yes, No, N/A
Churn management	If the customer requested a cancellation, did the agent make an effort to retain their business?	Yes, No, N/A
Closing	Did the agent ask for the customer’s permission before transferring the call? Did the agent ask if there was anything else the customer needed help with before closing the conversation? Did the agent encourage the customer to participate in the post-interaction survey? Did the agent recap the outcome of the conversation to ensure the customer's understanding? Did the agent set clear expectations for the next steps or follow-up actions?	Yes, No, N/A
Closing	Did the agent provide a polite and courteous goodbye message?	Yes, No
Custom	If your organization has an Analytics Enterprise Plus license, you can also add your own custom questions to an evaluation form. For more information about custom questions, go to the “Create a custom question” procedure in the Manage forms for AI evaluation article. If the custom question is based on custom metadata, please allow at least 45 minutes for new custom metadata fields to be available on the Auto QM Evaluation Form Manager page.
Custom Metadata	If your organization has configured custom metadata on the Metadata Manager page, you can create questions that Auto QM answers based on these metadata fields. If you add a new custom metadata field, please allow at least 45 minutes for the field to be available on the Auto QM Evaluation Form Manager page. For more information about custom metadata, see Manage custom metadata fields. For more information about writing questions based on custom metadata, see Advice for writing AI evaluation forms.
Customer effort	Did the customer express frustration with the agent? Did the customer request an escalation to someone in a higher position? Was the customer put on hold for more times than permitted? Did the call time exceed the call duration limit that is acceptable to a customer? Did the customer experience silence on the call for longer than the acceptable total silence time? Was the customer put on hold for longer than the acceptable hold time limit? (AI Agent only) Did the customer provide negative feedback in the conversation? (AI Agent only) Did the customer request to speak to a human agent more than once? (AI Agent only) Did the customer abandon the conversation before getting a resolution? (AI Agent only) Did the AI Agent provide sufficient content or guidance before the customer abandoned, so their needs were reasonably addressed? NOTE For further information about the acceptable number of holds, call duration, and so on, see Thresholds.	Yes, No, N/A
Customer effort	Did the customer have to repeat themselves to convey their issue or provide the same information to the agent multiple times?	Yes, No
Empathy	Did the agent avoid dismissing the customer’s emotions or complaints? Did the agent express empathy at the right time and appropriate to the emotions shared by the customer? Did the agent express empathy and provide reassurance to the customer?	Yes, No, N/A
Greeting	Did the agent ask about how the caller heard about the company or service? Did the agent clarify whether the caller was calling for themselves or on behalf of another person? Did the agent confirm the best contact number to use in case of a disconnect?	Yes, No, N/A
Greeting	Did the agent clearly state their name and where they are calling from? Did the agent ensure they were speaking with the correct person before proceeding? Did the agent establish the purpose of the call clearly and concisely? Did the agent inform the customer that the call is being recorded? Did the agent personalize the conversation by introducing themselves with their name? Did the agent use a polite greeting when starting the conversation?	Yes, No
Intelligent phrase matching	Intelligent phrase matching lets you provide examples of specific phrases that should affect the evaluation score if they appear in the contact. For information about intelligent phrase matching, go to Manage forms for AI evaluation.
Objection handling	Did the agent effectively address any objections or concerns raised by the customer?	Yes, No, N/A
Professionalism	Did the agent address the customer by name at least once during the conversation? Did the agent avoid displaying a negative attitude which can consist of sarcasm, frustration, annoyance, or any other negative emotions? Did the agent avoid using profanity or inappropriate language? Did the agent avoid using unnecessary jargon when providing support or explaining the situation to the customer? Did the agent demonstrate flexibility by adapting to the customer's needs and exploring alternative solutions when necessary? Did the agent maintain a positive and optimistic attitude when providing information and solutioning during the conversation? Was the agent courteous and polite throughout the conversation? Did the agent avoid talk-overs? (AI Agent only) Did the AI Agent adapt its language level appropriately to the customer’s knowledge and tone?	Yes, No, N/A
Resolution	Did the customer demonstrate satisfaction with the resolution, either through explicit statements, affirmative responses, or by indicating they can proceed without further assistance? Did the agent accurately identify and address the primary issue or query raised by the customer, demonstrating a clear understanding of the problem? Did the agent actively lead the resolution process by proposing a solution, clarifying doubts, or answering the customer's inquiry? (AI Agent only) Did the AI Agent escalate when it was necessary? (AI Agent only) Did the AI Agent fully resolve the customer’s request or question without escalation?	Yes, No, N/A
Sales	Did the agent correctly identify an opportunity and attempt to promote relevant products or services from the company? Did the agent effectively address the customer's sales objections? If this is a sales call, was the agent able to introduce the product/service/offer in a concise and engaging way?	Yes, No, N/A

Thresholds

Webex WFO currently presets the following thresholds in the standard evaluation form.

Individual hold time - more than 2 minutes
Number of hold events - more than 3 holds
Total silence time - more 10 seconds (exclude silences of less than 2 seconds)
Conversation duration - more than 6 minutes

NOTE These thresholds are also configurable in the Auto QM Evaluation Form Manager. For more information, see Manage forms for AI evaluation.