---
sourceDocument: Yokohama Enable AI
sourceDocumentLink: https://servicenow-prod.fluidtopics.net/r/yokohama/intelligent-experiences

 Release :

    - yokohama

ft:locale :

    - en-US

ft:publication_title :

    - Yokohama Enable AI

ft:clusterId :

    - platai

bundleId :

    - platai

workflow :

    - Platform


---

# Human feedback for evaluations

# Human feedback for evaluations {#ariaid-title1}

Release version: Yokohama  
Updated August 13, 2025  
![](https://www.servicenow.com/docs/portal-asset/ico-clock) 1 minute to read  
Expand the Human feedback section to see details on evaluations and their satisfaction scores.
Role required: interaction_admin  
Note:  
To label a record or label a conversation, you must have access to the transcript, and only users with the interaction_admin role have access to the transcript.

The Evaluations section shows all the chats that were auto-evaluated by the large language model (LLM). You have the option to manually evaluate conversations to compare the AI evaluations against your own interpretation of how the
agent conversation was.  
{#ai-eval-human-fb__table_cvn_fjb_yfc__entry__2}

| Field | Description |
|-|-|
| Number | Evaluation number assigned to each chat conversation. Select the evaluation number to see the respective chat and its evaluations. |
| State | State of the evaluation. |
| Auto Eval user satisfaction Score | Satisfaction score calculated automatically by the LLM. |
| Human User Satisfaction Score | Satisfaction score calculated based on the user evaluation of the chat. |
| Gap | Difference between the human and auto-evaluated satisfaction scores. |
[ ]

{#ai-eval-human-fb__table_cvn_fjb_yfc}

## Manually evaluate a chat {#ai-eval-human-fb__section_yp5_p44_xfc}

You can manually evaluate each chat to compare it with the AI evaluation.

1. Select the evaluation that you want to score manually.
2. Toggle the View Auto Eval Scores switch on to see the AI evaluation for each category.
3. For each category, select your answer for how accurately you feel that the agent responded.

4. Toggle the Other Metrics switch on for a more detailed evaluation.
5. After completing, select Submit.
{#ai-eval-human-fb__ol_jpq_1p4_xfc}

The Human user Satisfaction Score value gets calculated based on your response to the questions. You can see your responses for each evaluation by selecting it and then selecting View Human Scores. The Export option enables you to export the data in your preferred format.

You can also randomly label evaluations by selecting Label random scores. When selected, a list of 10 random, unevaluated conversations from the past 10 days are loaded for manual labeling.