{"revision":"c10276e953456835cdbb1f3a60d58ae1980be353","document":{"id":7015,"path":"customer-service-management/now-assist-for-csm/sentiment-analysis-interaction.md","title":"Sentiment analysis for email interactions","bundle":null,"doc_type":null,"release":"australia","body":"\n# Sentiment analysis for email interactions\n\nConfigure Sentiment analysis for email interactions to evaluate sentiment, trends, and reasoning at interaction level and help agents resolve cases more efficiently.\n\n## Before you begin\n\nRole required: admin\n\nInstall the plugin \\(sn\\_eaai\\_csm\\) for Email Interaction for CSM.\n\n## About this task\n\nSentiment analysis for email interactions is generated from the information that you enter in the following fields:\n\n-   Description\n-   Short Description\n\nAny modifications to the names or labels of these fields can result in issues with sentiment analysis.\n\nSentiment is calculated using several factors:\n\n-   Latest requester message\n-   Overall conversation between agent and customer captured in transcript\n-   Sentiment history \\(trend\\)\n-   Issue complexity\n-   Feedback or continued frustration at the end of conversation\n-   Frequency of updates and response time\n-   Language tone\n\nThe Sentiment Reasoning field provides details on which factors were considered. This includes:\n\n-   Tone of the requester's messages across the transcript.\n-   Feedback or continued frustration in the latest requester's messages.\n-   Complexity of the issue based on interactions from both the requester and the agent.\n\n## Procedure\n\n1.  Navigate to **Admin &gt; AI Admin Hub &gt; AI Skills**.\n\n2.  Select the **Customer** workflow, and **CSM** as the product.\n\n3.  Activate skill for the **Sentiment analysis for email interactions** skill.\n\n    Each skill has a guided setup with multiple steps. A check symbol next to each step indicates whether its setup is complete, partially complete, or incomplete. After configuring a step, select **Save and continue** to move forward, or **Back** to return to a previous step.\n\n4.  Select **General details** and review the name and description of the skill.\n\n    Additional information regarding details of the skill are displayed, but can't be edited.\n\n5.  Select **Choose Input** and review the tables and fields to create prompts that determine where data is pulled from.\n\n    **Note:** You cannot modify the input data source.\n\n<table id=\"table_qrf_yz5_r3c\"><thead><tr><th>\n\nInput\n\n</th><th>\n\nDescription\n\n</th></tr></thead><tbody><tr><td>\n\nInput table\n\n</td><td>\n\nInteraction\n\n</td></tr><tr><td>\n\nInput fields\n\n</td><td>\n\n-   Description\n-   Short Description\n\n\n</td></tr></tbody>\n</table>6.  Select **Role attribution**\n\n    **Note:** You cannot modify the input data source in base system.\n\n<table id=\"table_dj3_lg5_r3c\"><thead><tr><th>\n\nInput\n\n</th><th>\n\nDescription\n\n</th></tr></thead><tbody><tr><td>\n\nInput table\n\n</td><td>\n\nInteraction\n\n</td></tr><tr><td>\n\nRequester fields\n\n</td><td>\n\n-   Opened for\n-   Created by\n\n\n</td></tr><tr><td>\n\nFulfiller fields\n\n</td><td>\n\nAssigned to\n\n</td></tr></tbody>\n</table>7.  Select **Define Availability** to customize how and when the skill capability is active and accessible.\n\n    -   Select **Skill is always available** so no restrictions are placed on when a skill is available.\n    -   Select **Customize skill availability** to define conditions and use the condition builder to configure fields and values.\n8.  Select **Define Trigger** and edit the job schedule frequency period to configure how often the skill will be triggered.\n\n    To enable a scheduled job for a custom table, follow these steps:\n\n    1.  Navigate to **All** &gt; **System Definition** &gt; **Scheduled Jobs**.\n    2.  Search and select **Sentiment analysis scheduler job interaction CSM**.\n    3.  Replace the table name **interaction** with the name of your custom table in the script.\n    4.  Select **Update**.\n9.  Select **Save and continue**.\n\n10. Select **Define access** to determine who can access this skill.\n\n    By selecting specific roles, you're controlling who can use it. The roles you choose will also be available in the next step **Select display**.\n\n    Default and Custom Roles:\n\n    -   If no changes are made, the default roles sn\\_customerservice\\_agent and sn\\_customerservice.consumer\\_agent will automatically appear in **Define Access** and **Select Display**.\n    -   If custom roles were added before the upgrade, they’ll be updated automatically by a script.\n    -   If new roles are created after the upgrade, you must manually add them in both the **Define Access** and **Select Display**.\n\n        **Note:** In the **Select Display** step, you can only choose roles that were added in the **Define Access** step. If you add a role in **Define Access**, you still must manually select it in **Select Display** to make it active.\n\n11. Select **Save and continue**.\n\n12. **Select display** and toggle to determine if the skill appears in In-product desktop, displaying Sentiment analysis for email interactions in all CSM products.\n\n13. Select **Save and continue**.\n\n14. After selecting **Review and Activate** to examine changes, select **Done** to close the sentiment analysis generation settings.\n\n15. Select **Activate** to turn on the skill for agents and complete the configuration.\n\n\n## What to do next\n\n-   AI Skill Kit is the ServiceNow framework for [cloning](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/australia/customer-service-management/now-assist-for-csm/clone-the-now-assist-for-csm-skills.md) and [customizing](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/australia/customer-service-management/now-assist-for-csm/customizing-now-assist-skills.md) AI skills. When configuring Sentiment Analysis for email interactions, the base table provides the core email record data. Additional inputs supply the LLM with broader context- such as email thread history, linked case fields, or customer contact data that can improve the accuracy of sentiment output.\n-   You can customize the email sentiment prompt. Follow the steps to customize:\n    1.  Access the OneExtend Capability Definition record by going to the sys\\_generative\\_ai\\_config table. Here is an example for Azure OpenAI \\(other providers like Claude and Gemini will have a different sys\\_id\\): https://anoway1.service-now.com/sys\\_generative\\_ai\\_config.do?sys\\_id=479cdabf93b48310003afba6dd03d606\n    2.  Go to the **Prompt Template** field and edit the prompt.\n    3.  **Save** the record.\n\n        **Note:** Verify that the user is in ServiceNow Otto for Customer Service Management \\(CSM\\) scope and the record is still in **Published** state.\n\n\n**Related topics**  \n\n\n[Analyze sentiments in ServiceNow Otto for Customer Service Management \\(CSM\\)](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/australia/customer-service-management/now-assist-for-csm/analyze-sentiments-in-now-assist-for-csm.md)\n\n[Enable and configure the Sentiment Analysis](https://support.servicenow.com/kb?sys_kb_id=ec531aab47a48794b6d8aa25126d4322&id=kb_article_view)\n\n"}}