Article
    by Pablo Larrosa, Vice President of Technology

    ServiceNow® AI blueprint: Issue Auto Resolution

    Issue Auto Resolution (IAR) helps organizations enable employees to put in a request in their own natural language and receive a fast response, avoiding triage bottlenecks.

    Service

    ServiceNow

    Last year, with the Xanadu release, ServiceNow® introduced several AI features that seamlessly integrate with the whole platform. From machine learning for automation intelligence to virtual agent natural language understanding (NLU). One of the standout capabilities is Issue Auto Resolution, a virtual agent designed to proactively resolve issues. 

    How does Issue Auto Resolution work?

    When a user creates an issue, the IAR application uses advanced machine learning and NLU to assess the issue's criticality and intent. Then, it selects the most appropriate virtual agent topic for automatic resolution.

    If an exact virtual agent topic isn't available, Issue Auto Resolution performs an AI search to find the most relevant knowledge articles, FAQ, or catalog content. The deflection information is then delivered to the user through their preferred communication channel, such as email, SMS, or virtual agent.

    User journey in Issue Auto Resolution

    1. Someone creates a ticket.

    2. The platform assigns the incident to a virtual agent.

    3.1. If a virtual agent finds a matching topic, it responds automatically.

    3.2. If a virtual agent topic isn’t found, IAR uses AI search for relevant articles, FAQ, or catalog content.

    Set the foundations for IAR

    To gain the best performance for any AI solution out there, you need to have quality data that the model can lean on—complete data, with no inconsistencies, misspellings, or incorrect values.

    To train the model for IAR, ServiceNow recommends at least 10,000 incidents. Another thing to take into consideration is that not every solution is available for every ServiceNow deployment architecture. For example, you can’t use IAR in a self-hosted environment; you need to enable it in a customer instance. None of the AI features can be enabled in a personal dev instance.

    To analyze if a feature could be beneficial, contact your vendor and ServiceNow sales rep to define the next steps.

    Key metrics for evaluating IAR impact

    This agent can proactively resolve issues for ​ITSM and HR, improve the mean time to resolution (MTTR), and improve employee experience. Another key metric that has a direct impact is the first contact resolution rate, which indicates the percentage of problems that are resolved by first-level IT support.

    It’s also good practice to identify the change in the number of tickets handled by the helpdesk team after enabling the agent.

    Roadblocks for implementation

    There are some scenarios that aren’t suited to this solution, such as ServiceNow instances that are heavily customized with several legacy systems. Before pursuing this AI tool, define your phase zero to identify your starting point and desired roadmap.  

    Security considerations

    IT handles vast amounts of data, from password reset requests to computer hardware issues. They also manage data related to applications, infrastructure, and services.

    The data assets stored in the configuration management database (CMDB) include information about the organization's internal infrastructure and IT-related assets, such as server models, OS versions, patch levels, dependencies, and IP addresses. This key configuration data could be easily used to identify vulnerabilities in a system.

    To stay proactive against attacks, detailed maps of corporate infrastructure should be considered sensitive information. For this reason, it is vital to properly protect all data stored in the CMDB at all times.

    The main challenge with virtual agents

    Nowadays, every platform brings its own out-of-the-box agent, like Salesforce Einstein Agent, Microsoft Copilot, and others. In our experience, not all of them are ideal for a production environment and can’t solve every need. You will need to decide if it’s worth building a homegrown AI solution for sentiment analysis or better to wait for the next version of the platform AI agent.

    We can assist in analyzing, comparing, and defining the roadmap. Taking productivity and ROI into account means that decision isn’t easy—let us help.

    Get in touch

    Nortal is a strategic innovation and technology company with an unparalleled track-record of delivering successful transformation projects over 20 years.