Creating and training solutions

  • Release version: Xanadu
  • Updated August 1, 2024
  • 2 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 Creating and training solutions

    ServiceNow's Predictive Intelligence (PI) frameworks enable customers to create and train machine-learning solutions that predict, recommend, and organize data outcomes tailored to their business needs. These solutions leverage historical data to automate categorization, recommendation, and pattern identification in records, enhancing operational efficiency.

    Show full answer Show less

    Key Features

    • Multiple Frameworks for Different Solution Types:
      • Classification Solutions: Automatically categorize and route work during record creation based on past data.
      • Similarity Solutions: Recommend resolutions by identifying similarities between new and existing records.
      • Clustering Solutions: Group similar records to detect patterns and major incidents.
      • Regression Solutions: Previously used to predict numeric values like resolution times; however, creating new regression solutions is no longer supported as of the Xanadu release, though existing ones can still be edited and trained.
    • Data Selection for Effective Training: To maximize prediction accuracy, training data should include:
      • Input fields available at record creation to enable real-time predictions.
      • Output fields restricted to a finite set of choice values for clarity.
      • Only records with reliable and correct output values, excluding recently changed or reviewable data.
      • Multiple examples covering all output field values and input field variations to ensure broad coverage.
    • Secure Training and Prediction Process:
      • Solutions and associated records are exported to a centralized training server within the same datacenter, ensuring compliance with data sovereignty requirements.
      • After training, the solution is returned to the instance, and training data is deleted from the server.
      • Predictions use a centralized prediction server in the same datacenter, caching trained model artifacts for efficiency.
      • All communications occur securely over HTTPS within the datacenter firewall.

    Next Steps and Support

    Customers can create and manage their solutions via the Predictive Intelligence homepage in their instance. For troubleshooting training issues, ServiceNow provides a dedicated knowledge base article addressing common problems.

    Use one of the Predictive Intelligence (PI) frameworks to create and train machine-learning solutions. Each framework delivers a different solution type for training the system to predict, recommend, and organize data outcomes.

    Types of solutions

    The three PI frameworks provide different solutions that can be invoked by any application through a prediction API to make a prediction. Create and train your own solutions using your previous data. Navigate to All > Predictive Intelligence > Homepage to view and create solutions.

    Select the best framework for your desired prediction:
    • Classification solutions:

      Sets field values during record creation to automatically categorize and route work based on past records. See Create and train a classification solution.

    • Similarity solutions:

      Identifies similarities between new and existing records to recommend resolutions. See Create and train a similarity solution.

    • Clustering solutions:

      Groups similar records into clusters to identify patterns and major incidents. See Create and train a clustering solution.

    • Regression solutions:
      Important:
      With the Xanadu release, support for creating new regression solutions was removed. You can still edit and train existing regression solutions, but you won't be able to create new ones.
      Uses historic data to predict numeric outputs, such as estimating the time it takes to resolve an incident or case. See Create and train a regression solution.

    Selecting data records for training your solution

    A solution is only as good as the record data you use to train it. In general, a good training dataset has these characteristics.
    • The solution definition input fields are available to users when creating records. To make predictions at record creation, the solution must have the input field values at record creation.
    • The solution definition output field is a choice field. To make more accurate predictions, limit the output field to a finite set of possible values.
    • The training records only contain correct values for the output field. To make more accurate predictions, filter out any records that have unreliable output field values. For example, if recently closed incidents are subject to review and change for a month, filter out any recently closed incidents.
    • The training records contain multiple examples of each output field value that you want the solution to predict. To provide more record coverage, include multiple examples of each output field value.
    • The training records include common variations of the input fields. To provide more record coverage, include multiple examples of input field values.

    Exporting your solution for training

    Predictive Intelligence training flow

    To train a solution, you export its solution definition and associated records to a centralized training server within the same datacenter. When the training completes, the training server exports the solution back to your instance and deletes all of your training data from the server. As every datacenter has its own dedicated training server and the data doesn't leave the datacenter, this service is also available to customers who have data sovereignty requirements.

    Predictions occur on a centralized prediction server within the same datacenter as the instance. The trained model artifacts are sent from the instance server to the prediction server when the prediction is invoked for the first time. After that, the trained model artifacts are cached on the prediction server for subsequent predictions.
    Note:
    All communication between the instance and the training service occurs within the same datacenter firewall. Even so, all communications occur over HTTPS.

    Solution training troubleshooting

    For troubleshooting common training issues, see the Predictive Intelligence Common issues [KB781893] article in the Now Support Knowledge Base.