--- sourceDocument: Brazil Customer Service Management sourceDocumentLink: https://www.servicenow.com/docs/r/customer-service-management Release : - brazil ft:locale : - en-US ft:publication_title : - Brazil Customer Service Management ft:clusterId : - csm bundleId : - csm workflow : - Customer and Industry --- # Live interaction recommendations AI agent # Customer Service Management Live interaction recommendations AI agent for voice channel {#ariaid-title1} Release version: Brazil Updated September 10, 2026 ![](https://www.servicenow.com/docs/portal-asset/ico-clock) 8 minutes to read Summarize ![AI sparkle icon](https://servicenow.com/docs/portal-asset/ai-sparkle-icon) Summarized using AI This content was generated using new OpenAI-powered functionality. Results are provided on an as is basis and are not guaranteed to be accurate or complete. ## Summary of Customer Service Management Live Interaction Recommendations AI Agent for Voice Channel The Live Interaction Recommendations AI agent in ServiceNow Otto® Customer Service Management (CSM) delivers real-time, AI-driven recommendations to human agents during live voice interactions. It analyzes conversation transcripts, leverages customer context and knowledge bases, and surfaces actionable guidance within the CSM Configurable workspace. This capability improves agent efficiency by reducing handle time, enhancing first-contact resolution, and ensuring consistent, accurate responses throughout calls. Show full answer Show less Starting with version 7.0, this AI agent was renamed from Live Agent Assist AI agent, along with some entity name changes. ## How It Works * **Get Interaction Context:** Automatically fetches baseline customer and interaction data (e.g., account info, sold products, prior cases) when a voice call is assigned, without manual agent input. * **Analyze Interaction \& Generate Recommendation:** Processes live conversation transcripts, generates Knowledge Graph queries, searches knowledge articles, and synthesizes cited, actionable recommendations for agents. The AI agent is triggered automatically upon call assignment in the CSM workspace, requiring no manual initiation. ## Key Skills * **Transcript Analyser:** Extracts unanswered customer queries and evaluates resolution options. * **Graph Query Generator:** Builds Knowledge Graph queries to retrieve integrated customer data. * **Response Generator:** Combines knowledge articles and customer context into final recommendations with source citations. * **Query Enhancer:** Enriches agent follow-up questions with context and transcript history for ongoing recommendation accuracy. ## Data Sources * **Customer Context Builder:** Pre-fetches fast-access customer data from multiple CSM tables (cases, contracts, assets, entitlements) via a scripted interface, supporting customized extensions. * **AI Search Profile:** Retrieves knowledge articles (default) and can be extended to additional sources for recommendations. * **Enterprise Knowledge Graph:** Provides targeted customer-specific data by querying a defined set of CSM-related tables, customizable by administrators. ## Security and Role Requirements Access to the AI agent's workflows requires specific roles such as **sncustomerserviceagent** for B2B and **sncustomerservice.consumeragent** for B2C agents. Administrators must configure role masking and security controls accordingly, including updating Access Control Lists (ACLs) to align with custom roles and ensure proper permissions for case resolution. ## Configuration and Activation * Enable the Customer Service Management AI Agent Collection plugin. * Activate the Live Interaction Recommendations AI agent and its interaction trigger within AI Agent Studio. * Adjust ACLs to permit necessary access aligned to your organizational security policies. ## Using the AI Agent in Voice Interactions Integration with a CCaaS provider and real-time transcription is required. When a customer call connects, an interaction record is created and assigned to an agent in the CSM workspace. Agents can open the Otto panel to view the interaction and select **Get Recommendations** to receive AI-generated suggestions based on live transcript analysis. The AI agent handles multiple customer queries by allowing agents to prioritize and address them sequentially. Recommendations come with source citations and can be expanded to show detailed knowledge articles. Additional features include: * **Clarification Handling:** The AI asks clarifying questions when needed to refine recommendations. * **Follow-up Query Handling:** Agents can ask natural language follow-up questions and receive contextually relevant answers that maintain conversational continuity. ## Benefits for ServiceNow Customers * Accelerates customer service by providing timely, context-aware recommendations during live calls. * Improves first-contact resolution rates and consistency of answers. * Enhances agent productivity by automating data retrieval and knowledge synthesis. * Supports customization and extension to fit specific business needs through configurable data sources and role management. Live interaction recommendations AI agent is an agentic capability within ServiceNow Otto® Customer Service Management (CSM) that provides real-time recommendations to human agents during live customer interactions on the voice channel. When a customer contacts a support agent, this AI agent analyzes the ongoing conversation transcript, retrieves relevant knowledge and customer context, and surfaces actionable recommendations directly within the CSM Configurable workspace. Important: Starting with v7.0 of Customer Service Management AI agent collection, the Live Agent Assist AI agent has been renamed to Live Interaction Recommendations AI agent. Some of its entities have also been renamed. For a list of old and new names, see [Renamed entities in Live Agent Assist](https://www.servicenow.com/docs/oD1WHGeBaAymX6BovDuPmg "Starting with version 7.0 of Customer Service Management AI agent collection, several configuration records that support the Live Agent Assist capability were renamed to align with its new label, Live interaction recommendations AI agent."). Live interaction recommendations AI agent reduces handle time, improves first-contact resolution, and helps agents deliver consistent, accurate responses throughout the live voice interaction. ## How Live interaction recommendations AI agent works {#csm-live-agent-assist__section-laa-overview} Live interaction recommendations AI agent consists of two tools: Get Interaction Context (Script and autonomous) : This tool retrieves contextual information about the current voice interaction. It is configured as a pre-run tool, so it executes automatically when a Live interaction recommendations AI agent is initialized. It gathers baseline customer and interaction details, such as account, install base, sold products, and previous cases or interactions, before any recommendation is generated. Because it is autonomous, it requires no manual action from the agent; it fetches this context proactively so that the Analyze Interaction \& Generate Response tool has the customer data it needs to process the customer query. Analyse Interaction and Generate Recommendation (Subflow and autonomous) : The tool is a subflow-based tool that internally invokes skills to analyze the transcript, generate Knowledge Graph queries and generate relevant responses. These skills work in sequence to process the conversation, retrieve information from multiple data sources, and generate a final recommendation with cited sources. Live interaction recommendations AI agent is triggered when the customer call lands to a human agent in the CSM Configurable workspace and a voice interaction record is created. It activates when a live CSM agent begins handling a customer interaction over a voice call. No manual initiation is required. Once triggered, Live interaction recommendations AI agent fetches basic customer information like account, install base, sold product details, previous cases, previous interactions. When the Get Recommendations button is selected, the Live interaction recommendations AI agent performs the following tasks: 1. Transcript Analysis: continuously reads and processes the live conversation transcript to extract unanswered customer queries. Based on the customer query: * Queries the Knowledge Graph if additional customer details are required. * Searches knowledge articles if recommendations need to be generated from a KB article. {#csm-live-agent-assist__ul_hsm_fh5_zjc}For more information on the data sources, see [Data sources](https://www.servicenow.com/docs/qDZMn4gdQqhVNTFWyRSRzQ#csm-live-agent-assist__laa-data-sources). 2. Recommendation Generation: synthesizes results for the customer query from the customer context, KB article, and Knowledge Graph to generate actionable, cited recommendations for the agent. 3. Clarification Query Handling: manages scenarios where the transcript is ambiguous, presenting clarifying questions to the agent. 4. Follow-up Query Handling: addresses agent follow-up questions in the context of the ongoing voice interaction. {#csm-live-agent-assist__ol_wg4_2g5_zjc} The following table shows the skills and their responsibility in the Live interaction recommendations AI agent: {#csm-live-agent-assist__table_cdq_c41_vjc__entry__2} | Skill | Responsibility | |-|-| | Live Interaction Transcript Analyser | Analyzes the live transcript to extract unanswered customer queries and evaluate resolution paths. | | Live Interaction Graph Query Generator | Generates Knowledge Graph queries to retrieve joined data across CSM entity types. | | Live Interaction Response Generator | Synthesizes KB articles, customer info, and KG results into a final recommendation with source citations. | | Live Interaction Query Enhancer | Enriches agent follow-up queries with customer context and transcript history before re-initiating the recommendation pipeline. | [Table 1. Live interaction recommendations AI agent skills] {#csm-live-agent-assist__table_cdq_c41_vjc} ## Data sources {#csm-live-agent-assist__laa-data-sources} Live interaction recommendations AI agent retrieves information from the following primary data sources: Customer Context Builder : The Customer Context Builder is an interface class that pre-fetches basic customer context directly from database tables. This direct query path provides faster access to high-priority fields such as recent interactions and installed products for the current caller. The base system includes the CustomerContextBuilder interface with a CustomerContextBuilderImpl implementation. This implementation fetches and aggregates customer context data from multiple ServiceNow tables, including interactions, cases, contracts, assets, and entitlements. All data is anchored to an interaction record and serialized into a pipe-delimited format suitable for LLM input. Note: The Customer Context Builder is a scripted customization. Extending it requires creating a custom implementation of the interface class. For most scenarios, adding tables to the Enterprise Knowledge Graph tag is an easier approach because it does not require scripting. AI Search Profile : A Retrieval-Augmented Generation (RAG) retrieval call queries the AI search profile to identify configured sources. In the base system, the Live interaction recommendations AI agent Search Profile is scoped to the Knowledge (kb_knowledge) table, so recommendations draw from knowledge articles. Source citations identifying the contributing articles appear at the bottom of each recommendation. Admins can extend the profile to include additional tables. For more information on how to add additional search sources, see [Create a search source for AI Search](https://www.servicenow.com/docs/r/platform-administration/ai-search/create-search-source-ais.html). Enterprise Knowledge Graph : The Enterprise Knowledge Graph retrieves customer-specific information such as contract details, recent cases, and asset data. A tag scopes queries to a defined set of tables so that retrieval stays targeted. In the base system, the tag covers the Account, Asset, Case, Consumer, Contact Contract, Entitlement, Install Base Item, Installed Product, Interaction, Product Model, and Sold Product tables. As an admin, you can edit the tag to include additional tables: 1. Navigate to AllKnowledge GraphKnowledge Graph DesignerEnterprise Graph. 2. From the Enterprise Graph Tags, select the CSM Live interaction recommendations AI agent Tag. 3. In the Selected Tables section, add the desired tables. {#csm-live-agent-assist__ol_tc3_w41_vjc} For more information on editing tags, see [Managing Knowledge Graph tags](https://www.servicenow.com/docs/access?context=managing-knowledge-graph-tags&version=brazil&pubname=brazil-intelligent-experiences&ft:locale=en-US). ## Role masking Required role: B2B agents (sn_customerservice_agent) and B2C agents (sn_customerservice.consumer_agent) Important: To access data in the agentic workflow, the admin role must include the specified roles under Contains roles. Agentic workflows and their AI agents use [role masking](https://www.servicenow.com/docs/access?context=aia-role-masking&version=brazil&pubname=brazil-intelligent-experiences&ft:locale=en-US) to determine which users can access them. Ones installed with your applications have specific roles that come included with the application. If you select Users with specific roles for user access, you must configure the security controls to include these roles. For the instructions to change the security controls, see [Define security controls for an agentic workflow](https://www.servicenow.com/docs/access?context=define-sec-controls-aw&version=brazil&pubname=brazil-intelligent-experiences&ft:locale=en-US). In the data access settings, you must also add the necessary roles to enable agents to resolve cases efficiently. For example, add the csm role to the agentic workflow's list of approved roles to enable access to case records. ## Configure Live interaction recommendations AI agent for voice channel {#csm-live-agent-assist__section_uvn_cbm_hhc} Configure the Live interaction recommendations AI agent to generate recommendations for ongoing voice interactions in the CSM Configurable workspace. Before you begin: * The Customer Service Management AI Agent Collection plugin must be activated on your instance. * Role required: now_assist_admin {#csm-live-agent-assist__ul_wkm_2p1_vjc} 1. Navigate to AllAI Agent StudioCreate and manage. 2. Select Live interaction recommendations AI agent in the AI agents tab. 3. Activate the trigger. * In the Add triggers section, select the Live interaction recommendations AI agent - Interaction trigger which is in inactive state by default. * In the Edit trigger dialog, turn on the Trigger is ON radio button and select Save. {#csm-live-agent-assist__ul_clw_kp1_vjc} For more information on modifying an AI agent, see [Modify an AI agent](https://www.servicenow.com/docs/access?context=modify-ai-agent&version=brazil&pubname=brazil-intelligent-experiences&ft:locale=en-US). {#csm-live-agent-assist__ol_vvn_cbm_hhc} ## Access Control lists (ACLs) {#csm-live-agent-assist__section_kj5_mks_fhc} ACLs (Access Control Lists) are preconfigured to support the AI agents and their associated flows and actions. By default, ACLs are configured for the sn_esm_agent role. Customers can modify these ACLs to align with their specific business requirements and security policies. For more information, [Configure security controls for a skill](https://www.servicenow.com/docs/access?context=nask-access-control&version=brazil&pubname=brazil-intelligent-experiences&ft:locale=en-US). When updating the agent role for the Live interaction recommendations AI agent, it is important to also update the corresponding ACLs to ensure proper permissions. To manually update ACLs for custom roles: 1. Go to the sys_security_acl table. 2. Use filters to locate ACLs related to your use case, AI agent, and internal flows or actions. 3. Add your custom role to each relevant ACL record. {#csm-live-agent-assist__ol_a2d_zhx_jgc} ## Use the Live interaction recommendations AI agent for voice channel {#csm-live-agent-assist__section_djt_5ks_fhc} Get real-time recommendations from the conversations happening over a live voice interaction. Before you begin: * ServiceNow and CCaaS provider integration must exist. * Real-time transcription feature with CCaaS provider must be enabled. * The Live interaction recommendations AI agent - Interaction trigger in the Live interaction recommendations AI agent must be activated. For more information, see [AI agents, skills, and agentic workflows on by default](https://www.servicenow.com/docs/access?context=now-assist-skills-on-by-default&version=brazil&pubname=brazil-intelligent-experiences&ft:locale=en-US). * Role required: sn_customerservice.consumer_agent {#csm-live-agent-assist__ul_erb_2q1_vjc} Procedure: 1. When a customer calls the support channel, an interaction record is created and assigned to a human agent. Select the ServiceNow Otto icon. The Otto panel appears with the interaction ID. 2. After you gather details about customer requirements, select the Get Recommendations button on the ServiceNow Otto panel. * The AI agent analyzes the conversation and generates recommendations or asks more questions. The processing icon appears while the agent processes the conversation before generating the next steps. If the customer asks multiple queries, you will be asked to choose the query you want to address first. You can pick a query and address one after another based on the customer's priority. For more information on how the recommendations are generated from the sources, see [Live interaction recommendations AI agent overview](https://www.servicenow.com/docs/qDZMn4gdQqhVNTFWyRSRzQ#csm-live-agent-assist__section-laa-overview). * When the recommendations are generated for a query, the sources from which the recommendations are generated appear under the Sources and more section. When you select the Sources and more button, the ServiceNow Otto panel expands where you can continue your conversation with the AI agent and view the list of sources simultaneously. You can open these sources for further details. {#csm-live-agent-assist__ul_y11_kq1_vjc} 3. As you progress through the conversation, select the Get Recommendations button to get the recommendations. 4. Clarification handling: If the AI Agent needs additional context before generating a recommendation, it displays clarifying questions with rendered option buttons, then waits for user selection. When you select one of the option buttons, the final answer is displayed with cited sources and the Get Recommendations button for further queries. 5. Follow-up query handling: After an initial recommendation is delivered, type any follow-up question in the text field and select the Enter icon to get the recommendations.You can ask natural language questions, such as What products is this customer currently using? The AI agent responds with accurate, contextually relevant answers and maintains conversational context across multiple questions. {#csm-live-agent-assist__ol_u11_kq1_vjc} The final response is rendered with cited sources and the Get Recommendations button. *[\>]: and then