Enterprise Architecture query agent

  • Release version: Australia
  • Updated May 26, 2026
  • 4 minutes to read
  • Summarize
    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 Enterprise Architecture query agent

    The Enterprise Architecture (EA) query agent is a Now Assist feature in ServiceNow that enables architects and portfolio managers to ask natural language questions about their enterprise architecture portfolio. It provides AI-generated, data-driven answers based on the organization's EA data without requiring manual report building or dashboard navigation. This read-only agent interprets various EA data components such as business applications, capabilities, value streams, technology standards, and information objects, delivering structured responses grounded in your ServiceNow instance’s data.

    Show full answer Show less

    Key Features

    • Natural Language Queries: Users can ask questions about individual records, apply multiple filters, compare trends across fiscal periods, or perform impact analysis scenarios like decommissioning applications.
    • Data Domains Covered: Includes business applications (ownership, lifecycle, risk scores), business capabilities, application-capability relationships, technical reference models (TRM), information objects, impact analysis, and technical debt.
    • AI-Powered Responses: Answers are generated by a large language model (LLM) subscribed through the ServiceNow platform, with responses varying depending on the selected LLM.
    • Now Assist Integration: The query agent is accessible via the Now Assist panel within the Enterprise Architecture Workspace, allowing real-time conversational interactions.
    • Knowledge Graph Dependency: The agent relies on the ServiceNow Knowledge Graph, which aggregates and updates CMDB data and relationships. Proper Knowledge Graph setup and system property enablement are prerequisites.

    Important Considerations

    • The agent is read-only; it cannot create, modify, or delete records.
    • Complex queries involving more than two relationship types may yield partial results and should be verified against source data.
    • Large-scale aggregations provide indicative figures; precise values should be confirmed within your CMDB.
    • The Knowledge Graph update process may take several days depending on data volume, during which certain query responses may be incomplete.
    • AI-generated responses are strictly based on your instance’s CMDB data; the accuracy depends on data quality and completeness.

    Practical Use and Access

    To use the EA query agent effectively, ensure that the Knowledge Graph system properties are enabled and that your CMDB data is accurate and up to date. Access the agent through the Now Assist panel in your Enterprise Architecture Workspace by clicking the Now Assist icon and starting a conversation. This tool streamlines exploration and analysis of your enterprise architecture portfolio, helping you gain insights, identify risks, track technical debt, and model impact scenarios efficiently.

    The Enterprise Architecture query agent answers natural language questions about your enterprise architecture portfolio, returning AI-generated answers grounded into their Enterprise Architecture data.

    Enterprise Architecture Workspace home page with the Now Assist panel open, displaying a prompt to enter a natural language question about enterprise architecture.
    Note:
    Depending on your license, you will have access to certain application features, generative AI skills, agentic workflows, and AI agents. For more information, see ServiceNow product tiers.
    Important:
    This Now Assist skill is turned on by default. The skill will be automatically available to appropriate role users for the application. For more information, see Now Assist skills, agents, and agentic workflows on by default.

    The Enterprise Architecture Explorer Query Agent is a Now Assist agent for Enterprise Architecture. It lets architects and portfolio managers query their EA data using natural language, without navigating dashboards or building reports manually.

    The agent interprets the structure of your EA portfolio, including business applications, capabilities, value streams, technology standards (TRM), and information objects, and returns structured answers based on your data. When processing a question, the agent searches for relevant data and actions before generating a response. After answering, the agent may display suggestions for related questions to help you explore further. Responses are generated by the large language model (LLM) that your organization has subscribed to on the ServiceNow® platform. The responses may also differ depending on the selected LLM.

    You can ask questions about individual records, apply multiple filter conditions, compare trends across fiscal periods, or model impact scenarios such as decommissioning an application.
    Note:
    The Enterprise Architecture query agent is read-only. It retrieves and analyzes your EA data but does not create, modify, or delete any records in your instance.

    What you can ask about

    The Enterprise Architecture Explorer Query Agent answers questions across the following EA data domains:

    Table 1. EA data domains supported by the Enterprise Architecture Explorer Query Agent
    Domain Example queries
    Business applications Ownership, lifecycle phase, application category, CSAT and technical risk scores, retirement candidates
    Business capabilities Hierarchy, scores, capability-to-application coverage
    App capability relationships Applications supporting a capability, applications with no capability assignments
    Trends Score comparisons for one or more applications
    TRM and technology Software versions, lifecycle phase
    Information objects Producing and consuming applications, orphaned objects
    Impact analysis Capability and information object impact of decommissioning an application
    Technical debt Applications flagged for technical debt

    Important information to derive the best results

    To get the most accurate responses from the agent, remember the following:
    • The agent works best for questions involving one or two relationship types. For queries with three or more relationships, for example, "list all child capabilities, their providing applications, the information objects those applications use, and the services they consume", it's recommended to check the results against your source records for completeness.
    • Large-scale aggregations — For counts and percentages across large data sets, treat results as indicative figures. For precise values, verify against your source records.
    • The agent queries your enterprise architecture data using the Knowledge Graph, which is updated through scheduled jobs. If new relationship types, tables, or custom fields are added to your instance, the Knowledge Graph does a scheduled refresh to generate updated descriptions. Some specialized or less common data points may not be fully represented yet, which could lead to partial responses for certain queries.
    • For questions about a specific record, include the record name or identifier in your query (for example, "Is Order Management a leaf capability?") to help the agent return the most relevant response.

    Access

    The Enterprise Architecture query agent is available through the Now Assist panel across Enterprise Architecture Workspace. Select the Now Assist icon to open the panel and start a conversation with the agent directly.

    The Enterprise Architecture Workspace home page with the Now Assist icon and Ask Now Assist icons highlighted.

    ServiceNow Knowledge Graph dependency

    The Enterprise Architecture query agent uses the ServiceNow Knowledge Graph to answer questions about CMDB relationships and infrastructure data. An administrator must enable the required Knowledge Graph system properties before the agent can access this data.

    After the system properties are enabled, Knowledge Graph runs scheduled jobs to process your CMDB data. The time required for initial processing varies by the volume of data in your instance and may take several days or longer. During this period, responses that rely on CMDB relationship data may be incomplete. Results improve progressively as processing continues. If the Knowledge Graph data processing is in progress, the Knowledge Graph Designer displays a notification. This means that the scheduled jobs required to set up the enterprise graph are still running. The Enterprise Architecture query agent is not available until this process is complete and the notification is no longer displayed.. To speed up the process, you can import enterprise graph descriptions from a sub-production instance where setup has already completed. For more information, see Initial setup for Enterprise Graph schema in production instance.

    Knowledge Graph Designer page showing a setup-in-progress notification banner indicating the Enterprise Graph is not yet available for queries.

    For enable the Knowledge Graph system properties, see Enable Knowledge Graph system properties for the Enterprise Architecture query agent.

    All data displayed by the Enterprise Architecture query agent is sourced directly from the ServiceNow® Knowledge Graph, which is built from and continuously updated by your instance's CMDB data. When you submit a question, the agent queries the Knowledge Graph to identify relevant records, relationships, and attributes before generating a response.

    This means the agent does not draw on general knowledge or external sources. Every answer is grounded in the data that exists in your instance — such as CI relationships, application ownership, capability mappings, and lifecycle information stored in your CMDB. If a record does not exist in your CMDB, the agent will not surface it.

    Note:
    The accuracy and completeness of the agent's responses depend on the quality and completeness of your CMDB data. Keeping your CMDB up to date ensures the most reliable answers.