--- sourceDocument: Yokohama Conversational Interfaces sourceDocumentLink: https://servicenow-prod.fluidtopics.net/r/yokohama/conversational-interfaces Release : - yokohama ft:locale : - en-US ft:publication_title : - Yokohama Conversational Interfaces ft:clusterId : - convint bundleId : - convint workflow : - Platform --- # Virtual Agent interaction records # Virtual Agent interaction records {#ariaid-title1} Release version: Yokohama Updated January 30, 2025 ![](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 Virtual Agent interaction records Every Virtual Agent conversation in ServiceNow is recorded as an interaction record in theInteractions \[interaction\]table. These records capture the full conversation, including all topic elements and any live agent transfers, across supported chat channels such as chat widgets, native mobile apps, and messaging integrations. The Interaction Management plugin powers this functionality and is shared across Virtual Agent, Agent Chat, and Connect Support. Show full answer Show less Interaction records enable administrators, Virtual Agent administrators, and support staff to review conversation details and analyze the types of interactions---whether fully Virtual Agent-driven or involving live agent handoffs. ## Key Features * **Conversation Transcript:** Logs the full dialog between requester and agent (virtual or live). Transcripts over 4000 characters are stored as attachments. * **Related Tasks:** Automatically links cases, incidents, or requests created or updated during the conversation to the interaction record, if topics are designed using recommended Virtual Agent Designer utilities and system methods. * **Interaction Logs:** Shows all topics run during the conversation, including setup and user-selected topics, with execution order. * **Key Fields:** Includes agent average response time, conversation state and state reason, sentiment analysis (if enabled), domain, assigned agent, timestamps, and flags indicating live agent involvement. * **State Reasons:** Detailed reasons explain why a conversation ended (e.g., user ended, system timeout, agent ended), differentiating between Virtual Agent only, Virtual Agent with Live Agent, and Live Agent only scenarios. * **Security Controls:** Conversation transcripts are viewable to all users by default; scoped ACLs and Restricted Caller Access can be configured to protect sensitive data. ## Practical Use for ServiceNow Customers ServiceNow customers can leverage the Interactions table to monitor and analyze Virtual Agent conversations comprehensively. The detailed fields and transcripts assist in troubleshooting, improving Virtual Agent topics, and understanding customer engagement patterns. By designing Virtual Agent topics with recommended utilities, customers ensure that relevant incident or request records are linked to the conversation history, enhancing traceability. Understanding the **State Reason** values helps in interpreting conversation closures, supporting reporting and operational insights. Additionally, security best practices allow customers to safeguard sensitive conversational data while maintaining accessibility for authorized users. ## Next Steps * Access and customize the Interactions list to review conversations by navigating to *interaction.list*. * Design Virtual Agent topics to integrate with incident or case management using the Action and Lookup utilities, and use `vaSystem.attachRecordToConversation` to link tasks properly. * Configure access controls if sensitive information is exchanged in conversations. * Use sentiment analysis data and state reasons for enhanced reporting and handoff logic between Virtual Agent and live agents. Each time a Virtual Agent conversation occurs, an interaction record captures the entire conversation in the Interactions \[interaction\] table. The record includes all topic elements used in the conversation, as well as live agent transfers. ## Interactions \[interaction\] table {#va-interactions__section_t3s_3rv_rjb} Each conversation in an instance generates an interaction record in the Interactions \[interactions\] table, which logs the conversation between a requester and Virtual Agent. The interaction records include conversations that occur across the supported chat channels: chat widget, native mobile app, and messaging integrations. Note: The Interactions table is part of the Interaction Management plugin (com.glide.interaction). Virtual Agent, Agent Chat, and Connect Support use this plugin. These interaction records provide administrators, Virtual Agent administrators, support agents, and support managers with a convenient way to review what happened during a conversation. The records also provide a quick way to see the type of conversations that occur: Virtual Agent, live agent transfers to Agent Chat, and live agent transfers to Connect Support. Figure 1. Interactions \[interaction\] table Note: Conversation transcripts are available only for Virtual Agent records that have Virtual Agent conversations and live agent transfers to Agent Chat. * When a conversation transcript for live chat exceeds 4000 characters, the transcript is handled as an attachment to the interaction record. * File attachments that the user uploads are recorded in the Attachments \[sys_attachment\] table, which corresponds to the interaction record. Attachments are automatically displayed in the activity stream. * For Virtual Agent conversations that have live agent transfers to Connect Support, only the Virtual Agent conversation is included in the interaction transcript. Live agent conversations in Connect Support are recorded in the activity stream. {#va-interactions__ul_ezq_5gh_1qb} ## Key fields in the Interactions table {#va-interactions__section_ams_5x1_twb} To view the Interactions table, navigate to All, and then enter interaction.list in the navigation filter. You can configure the fields displayed in the table using the [list collector](https://www.servicenow.com/docs/access?context=t_UseSlushbuckets&version=yokohama&pubname=yokohama-platform-user-interface&ft:locale=en-US) (slushbucket). {#va-interactions__table_xs3_3hy_yjb__entry__2} | Field | Description | |-|-| | Agent average response time | Average time based on how long the agent takes to respond to each message sent by a requester. For example: ``` 00:00 Requester: I have a question. 00:01 Agent: What is it? 00:04 Requester: How do I reset my password? 00:05 Requester: I forgot it and can't access my account. 00:07 Agent: Here is the answer. ``` Pair #1 00:01-00:00 = 00:01 Pair #2 00:07-00:04 = 00:03 (00:01 + 00:03) / 2 = 00:02 average | | Agent chat | Flag that indicates the conversation was transferred to a live agent in Agent Chat: true or false. | | Application | Scope of the application. Scope determines the accessible files and data available to other applications. | | Assigned to | Name of the fulfiller: Live agent name or Virtual Agent | | Domain | The domain the conversation took place in. For more information, see [Domain separation and Virtual Agent](https://servicenow-prod.fluidtopics.net/E1j8F9a6Fe4dLE0drB0a_w "Domain separation is supported in the Virtual Agent application. Domain separation enables you to separate data, processes, and administrative tasks into logical groupings called domains. You can control several aspects of this separation, including which users can see and access data."). | | Number | Number of the interaction record. | | Opened | Date and time that the conversation started. | | Opened for | Name of the requester. Anonymous users are identified as Guests. | | Sentiment | Shows the sentiment level of the interaction, if sentiment analysis is enabled on the instance. This value can be shown to agents in the Agent Chat interface. Business rules can be created based on this value at handoff (Virtual Agent to live agent, live agent to live agent, or live agent to queue). Valid values are as follows: * Positive * Neutral * Negative * null (no value assessed) {#va-interactions__abcdef_ul_eqq_fgc_fsb} This value is also entered in the [Interaction Insights table](https://servicenow-prod.fluidtopics.net/YoLPF_1MtEg6iQhykD1_eQ "User sentiment analysis prepares agents for the tone of incoming conversations. Business rules can be created based on this value and agents can take appropriate actions on the chat, whether it comes from Virtual Agent, another live agent, or from a queue."). | | State | Conversation states: * In Progress: Conversation between requester and virtual agent is in progress. * Closed Complete: Virtual agent ends the conversation. * Closed Abandoned: Requester ends the conversation before completing the conversation with the virtual agent. {#va-interactions__ul_pwr_fny_yjb} Note: There are different State Reasons for the Closed Complete and Closed Abandoned states, depending on whether you're using Virtual Agent only, Virtual Agent with Live Agent, or Live Agent only. For details on the State Reasons, see [Reasons for Closed Complete and Closed Abandoned states](https://servicenow-prod.fluidtopics.net/ZuuRpg0W0qr6qLZSr9CEqA#va-interactions__section_vxj_vcq_b4b). | | State Reason | Detailed explanations (reasons) for the conversation states Closed Complete and Closed Abandoned. To learn more, see [Reasons for Closed Complete and Closed Abandoned states](https://servicenow-prod.fluidtopics.net/ZuuRpg0W0qr6qLZSr9CEqA#va-interactions__section_vxj_vcq_b4b). | | Transcript | Log of the conversation between the requester and agents (virtual or live). | | Type | Type of interaction record. Virtual Agent conversations are record type Chat. | | Updated | Date and time that the conversation record was last updated. | | Virtual agent | Flag that indicates whether the conversation is for Virtual Agent: true or false. A Virtual Agent conversation is flagged when a requester selects a conversation topic or when the topic discovery process starts the appropriate conversation. | [ ] {#va-interactions__table_xs3_3hy_yjb} ## Reasons for Closed Complete and Closed Abandoned states {#va-interactions__section_vxj_vcq_b4b} Virtual Agent closes completed and abandoned conversations for various reasons, depending on whether the conversation is for Virtual Agent (VA) only, Virtual Agent with Live Agent (VA-LA), or Live Agent (LA). Figure 2. Interactions table with State Reason column The State Reason field provides details on the different ways in which a conversation ends. The following table describes the reasons and also the related message displayed for dashboard analytics. For details about how Virtual Agent and Agent Chat handle idle conversations and the corresponding default timeout periods, see [Closing Virtual Agent and Agent Chat conversations](https://servicenow-prod.fluidtopics.net/mqMgKoNj8lMwB7caplYo9w "Virtual Agent and Agent Chat conversations that are abandoned by requesters remain open until they are automatically closed by the scheduled job, Time Out Abandoned VA Conversations. This job runs hourly each day.").{#va-interactions__table_wxj_vcq_b4b__entry__4} | State | State Reason | Virtual Agent | Virtual Agent \& Live Agent or Live Agent only | |-|-|-|-| | Closed Abandoned | Left Before Engagement | Not applicable. | Requester chose to be routed to a live agent, but closed the conversation (selecting + End) before an agent was assigned. Dashboard message: User closed LA - before Agent engagement | | Closed Abandoned | No Activity | Requester opened a conversation, but didn't engage with the bot (for example, closed the browser) and didn't initiate topic discovery. The system closed the Virtual Agent idle conversation after the default timeout. * Virtual Agent idle chat timeout: Two hours (default) * Dashboard message: System closed VA - user never engaged {#va-interactions__ul_ygk_nqq_b4b} | Requester chose to be routed to a live agent (not through pre-chat setup) but didn't engage with the agent. If the conversation timed out and the Time Out Abandoned VA Conversationsscheduled job ran, the interaction state is Closed Abandoned with No Activity. Dashboard message: System closed LA - before Agent engagement | | Closed Complete | User Ended | After engaging with the bot (requester entered at least one utterance during a conversation or completed a conversation), requester closed the conversation session by selecting End. Dashboard message: User closed VA - Clicked End/X | Requester started to chat with live agent but closed the chat client by selecting End. Dashboard message: User closed LA - Clicked End/X | | Closed Complete | Auto Closed | Requester interacts with VA topic and completes the conversation. No live agent interaction. Dashboard message: System closed VA - Auto Closed | Not applicable. | | Closed Complete | No User Response | Requester engaged with the bot (entered at least one utterance), but left without actively closing the conversation. The system closes the idle conversation after two hours (default). * Virtual Agent idle chat closed: Two hours * Dashboard message: System closed VA - User no response {#va-interactions__ul_tsb_5yq_b4b} | Requester engaged with live agent but stopped and left the agent hanging (idle live chat). The system closes the idle live chat after six minutes (default). * Idle live chat closed: Six minutes * Dashboard message: System closed LA - User no response {#va-interactions__ul_zcr_4cr_b4b} | | Closed Complete | Topic Complete | Requester engaged with the bot in a topic but didn't respond to the Anything Else setup topic, or chose Yes to continue but didn't respond. The Time Out Abandoned VA Conversations job closes the interaction after the default timeout. * Virtual Agent idle chat closed: Two hours * Dashboard message: System closed VA - Topic complete {#va-interactions__ul_wrw_1dr_b4b} | Not applicable. | | Closed Complete | Left for search | Requester engaged with the bot (entered at least one utterance), but selected Search in the middle of the conversation and didn't return to the conversation. The system closes the conversation because of inactivity. Dashboard message: System closed VA- Left with AI search | Not applicable. | | Closed Complete | Agent Ended | Not applicable. | After requester started a chat with a live agent, the agent actively closed the conversation by selecting End Chat. Dashboard message: Agent closed LA - Clicked End/x | | Faulted | Error | A runtime exception occurred while processing a message. | Requester's conversation was closed due to a runtime error. | [ ] {#va-interactions__table_wxj_vcq_b4b} ## Detailed view of a Virtual Agent interaction record {#va-interactions__section_jp5_3f1_zjb} From the Interactions table, you can open a specific record to see additional details about the Virtual Agent interaction: a transcript of the conversation, related tasks, and interaction logs. Note: If a field is not visible, you can add it to the form view. For more information, see Configuring the form layout. Conversation transcript : The transcript shows the dialog exchange between the agent (virtual or live) and requester, as well as the time stamp for each dialog turn. Related Tasks : Any cases, incidents, or requests created or updated during the conversation are automatically attached to the interaction record as a Related Task. To ensure that these tasks are included with the interaction record, design your conversation topics using certain Virtual Agent Designer utilities and the system method vaSystem.attachRecordToConversation. For details, see [Topic design considerations](https://servicenow-prod.fluidtopics.net/ZuuRpg0W0qr6qLZSr9CEqA#va-interactions__section_nd1_fs1_zjb) below. Interaction Logs : The Interaction Logs tab lists all the topics that run in a conversation, including the setup topics and the topics executed by topic discovery or by user selection. Each topic element is assigned a number that indicates the sequential order in which the topics run. ## Security considerations {#va-interactions__section_d4n_bn1_zjb} Information in conversation transcripts is viewable to all users. If you need to control access to records containing sensitive information, you can use scoped access control lists (ACLs) and lock scripted access to those records by using the Restricted Caller Access (RCA) feature. For details, see [Set application scope, application resource, and event access](https://www.servicenow.com/docs/access?context=scope-resource-access&version=yokohama&pubname=yokohama-application-development&ft:locale=en-US) and [Manage Restricted Caller Access](https://www.servicenow.com/docs/access?context=t_ConfigureSIM&version=yokohama&pubname=yokohama-security-management§ion=restricted-caller-access&ft:locale=en-US). ## Topic design considerations {#va-interactions__section_nd1_fs1_zjb} If your Virtual Agent conversations create or update incidents, cases, or requests, consider designing your topics with the following [Virtual Agent Designer utilities](https://servicenow-prod.fluidtopics.net/XlcbYdmDY_STaUQ8Uuwnlw "You can control interactions within a topic by using utilities in Virtual Agent Designer. These utilities perform various actions such as adding topic blocks (reusable components), running a script, or adding different conversation paths (branches) in a topic. Utilities are represented as nodes in a conversation flow but do not display anything to the user.") and system method: * [Action utility](https://servicenow-prod.fluidtopics.net/wA0L2dM3~Em5QA~xiBtHBQ "Use the Action utility in a Virtual Agent topic to select the flow action properties that the node should create. The Action utility integrates Virtual Agent with the ServiceNow Workflow Studio product.") to create or update incidents or cases * [Lookup utility](https://servicenow-prod.fluidtopics.net/jNKMBkCMcfjpKIwhILiXAQ "Use the Lookup utility in a Virtual Agent topic to return a ServiceNow record query.") to view information * System method vaSystem.attachRecordToConversation(String tableName, String sysId) to attach the task to the interaction record {#va-interactions__ul_nhx_11b_zjb}Use these utilities and system method so that incidents, cases, or requests created during a Virtual Agent conversation are included in the Related Tasks list of the interaction record. Note: Records created or updated using the [Script Action utility](https://servicenow-prod.fluidtopics.net/kjZBpkuAR8qQfvY_53JgeA "Use the Script Action utility in a Virtual Agent topic to run a script.") are not captured as related tasks in the interaction record. **Related concepts** * [Domain separation and Virtual Agent](https://servicenow-prod.fluidtopics.net/E1j8F9a6Fe4dLE0drB0a_w "Domain separation is supported in the Virtual Agent application. Domain separation enables you to separate data, processes, and administrative tasks into logical groupings called domains. You can control several aspects of this separation, including which users can see and access data.") * [Virtual Agent scripts](https://servicenow-prod.fluidtopics.net/ke~GUHWgbfbG1gVpGBXv6A "Use ServiceNow Virtual Agent script methods and variables to write chat scripts, such as response, trigger, and flow scripts. Variables can also provide context for your live support topics and conversations.") **Related reference** * [Virtual Agent roles](https://servicenow-prod.fluidtopics.net/ADLLISrHgiW5P15~sMGPeQ#VA-roles "Virtual Agent adds several roles for users operating as Virtual Agent administrators.") * [Input data types in Virtual Agent topics](https://servicenow-prod.fluidtopics.net/ez3sicwydv8MfCVtr3s8~g "You can define inputs of commonly used Glide Virtual Agent (com.glide.cs.chatbot) data types in Virtual Agent Designer without writing a script. Define the input data type on the Start node of a custom control or topic block.") * [NLU system entities](https://servicenow-prod.fluidtopics.net/383zlgZc~i_gJ7dA3DMuXg "Use globally defined NLU entities to identify system information that Virtual Agent can extract from the conversation. You can define entities as \"nodeless\" input variables for a topic. These variables can be slot-filled from NLU service provider predictions or provided outside of the scope of the topic.") * [Virtual Agent URL parameters](https://servicenow-prod.fluidtopics.net/vNKDkTdIOc8plnioeUmuqQ "Virtual Agent provides various system parameters that admins can add to an instance URL to control how page content is rendered.") * [Latency feedback in Virtual Agent](https://servicenow-prod.fluidtopics.net/f_41JtGwhllDIecVv95iBg "The com.glide.cs.message.processing.enabled system property notifies requesters whenever the generative AI large language model (LLM) is processing their request in the Virtual Agent chat widget and Now Assist panel.")