Predictive Intelligence

Predictive Intelligence activates the most recent version of the solution when it completes training a solution. However, you can activate any previously trained solution version. Only one solution version can be active at a time, and only the active version is used when making predictions.

Analyze a cluster by a field available on the source table. With the Cluster Insight check box, you can add a filter condition on your input field when you review the list of results.

Apply purity to learn details about the composition of each cluster. For example, see what percentage of incidents in a cluster have the same value for Assignment Group. You can specify which fields to focus on, or you can let auto-purity display fields by default.

Tune the performance of a trained classification solution by configuring class level precision and coverage values.

Create and apply a class recall parameter to an ML solution prior to training its data. For example, you set and apply this solution parameter to 90% recall for all records in the Email class.

Name your clusters to help identify and organize them.

Apply Configure Connect Component and Levenshtein Distance method encoding to optimize the training for your clustering solutions.

Consider applying the Density Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm to your clustering solution. DBSCAN is available as an alternative to the default clustering algorithm, k-means.

Consider applying the Hierarchical Density Based Spatial Clustering of Applications with Noise (HDBSCAN) algorithm to your clustering solution. HDBSCAN is available as an alternative to the default clustering algorithm, k-means.

Configure solution definitions for Predictive Intelligence to train, test, and monitor solution definitions

Apply Term Frequency–Inverse Document Frequency (TF-IDF) encoding to classification, clustering, or similarity solutions for Predictive Intelligence.

Apply XGBoost encoding to optimize the training for your classification or regression solutions.

Learn about advanced settings for your Predictive Intelligence machine learning (ML) solutions. Apply optional technology and algorithms for classification, clustering, similarity, and regression capabilities.

Set values for precision, coverage, and recall statistics for a trained machine learning solution.

Group similar records into clusters so you can address them collectively or identify patterns.

Exclude common words you want the system to ignore during training and prediction.

Regression solutions have been deprecated. They could be used to predict numeric outputs, such as a temperature or a stock price.

Create and train a machine learning solution to collect and compare your existing records to new similar records. For example, you can compare the text in an open Incident record to a resolved Incident record to reuse its resolution.

Specify the records used to train a classification solution, what fields trigger a prediction, and how often you want to retrain your solution.

Build a collection of words and phrases that functions as the vocabulary the system uses to compare your instance records based on their textual similarity. You can think of the word corpus as a dictionary that you want your machine-learning system to understand.

Use database views to join two or more tables as input for your Machine Learning (ML) solution.

Domain separation is supported in the Predictive Intelligence 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.

Exclude a class from prediction if its precision or coverage aren't satisfactory. Excluding a class prevents the classification model from predicting a particular output field value.

Exclude a class from solution training to prevent the model from ever making predictions for a particular output field class. For example, you can exclude a particular incident category from training if you plan to retire or change the category.

View the top 25 most representative records of a cluster.

Refine and test your ML solutions iteratively on a non-production instance, and then use update sets to export the changes to your production instance. This practice mitigates the risk of retraining solutions on your live production instance.

Activate Predictive Intelligence on your instance and get started with basic configuration.

The Observability Dashboard offers a unified view and actionable insights for errors detected in Predictive Intelligence. Use this dashboard to visualize logged errors and gain information on prediction reliability and potential problem areas.

Customize the training behavior of your classification solution by passing in a dictionary of parameters in JSON format.

Analyze the importance of each input field to your model's predictions using model explainability. Create a Workflow Classification model that includes a graphical analysis of feature importance by executing the provided script.

Enable the LightGBM (Light Gradient-Boosting Machine) algorithm as an advanced option for training classification models.

Set a threshold for the minimum number of records a label must have in your dataset to be included in model training.

Limit your classification model to use only the most common classes when training. You can specify the number of classes to use.

Reduce noise in your classification model and enhance predictive accuracy by removing records with the label "others" from training data. These are records with a distribution frequency of under one percent.

Understand the reasons your machine learning solution made a specific prediction with Test time explainability, which returns a list of input features that influenced the prediction.

The Predictive Intelligence Usage Analytics dashboard is a central location to understand the effectiveness and overall value of all your Predictive Intelligence solutions. View metrics for model training successes and failures. Monitor prediction statistics including breakdowns by individual model.

Learn which types of encryption are supported for training Predictive Intelligence solutions.

Predictive Intelligence provides three different model frameworks in the Brazil release: classification, similarity, and clustering. Each framework specializes in different types of predictions.

Predictive Intelligence is a powerful interface to train machine learning models. With Predictive Intelligence, you can improve performance, efficiency, and flexibility to your systems across multiple business units.

Predictive Intelligence provides international language support. Learn which languages are available to Predictive Intelligence solutions.

The properties for Predictive Intelligence control certain parameters of its machine-learning solutions, solution training process, and caching.

Predictive Intelligence is installed with these roles.

ServiceNow Predictive Intelligence is a platform function that provides a layer of artificial intelligence that empowers features and capabilities across ServiceNow applications to provide better work experiences.

Save your trained machine-learning (ML) solution data during a system clone.

Validate that Predictive Intelligence still works after you make any configuration change such as applying an upgrade or developing an application. Copy and customize these quick start tests to pass when using your instance-specific data.

Reference pages for components of Predictive Intelligence.

Use the Prediction Results dashboard to determine if classification solution predictions are improving over time. Identify solutions that need refining or retraining.

Review the similarity examples generated during solution training to determine whether the similarity score threshold meets your business requirements.

The Solution Statistics dashboard in Predictive Intelligence has been deprecated in the Xanadu release. It provided precision and coverage statistics for each class in a classification solution.

Use the ML Solutions (ML Training Jobs view) module to monitor the training status and progress for Predictive Intelligence solutions. The module displays training jobs for both user interface and API solutions.

Learn about ServiceNow applications and features that leverage Predictive Intelligence. Solutions that you can adapt are available for various business units and industries.

Once your machine-learning (ML) solutions are trained, you can call on the Predictive Intelligence API to make a solution prediction. In this example procedure, we use the REST API Explorer application to test a similarity solution prediction for resolved incident recommendations.

Once your machine-learning (ML) solutions are trained, you can call on the Predictive Intelligence API to make a solution prediction. In this example procedure, we use the REST API Explorer to test a classification solution prediction for incident categorization.

Evaluate the coverage and precision of your machine-learning (ML) solutions by testing them. Once deployed, track their performance over time. Improve predictions by using performance information to refine your solutions.

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.

If you encounter issues during your solution training and solution prediction, follow these suggested resolutions.

After you review the similarity examples provided by the system, update your solution similarity score threshold if you want the results returned by the solution to be more or less similar.

Use Predictive Intelligence actions in Workflow Studio to create flows that incorporate your model predictions.

Use APIs to simultaneously submit multiple classification solutions for training based on the Group By field.

MLSolutionFactory scriptable objects enable defining ML functionality. You can use the APIs to compose data-driven functionality, such as subclustering large clusters or clusters with multiple PRBs attached.

Use ServiceNow Machine Learning (ML) APIs to train Machine Learning models and run inferences.

Train and use Predictive Intelligence solutions to accomplish various tasks.

View your solution training progress or statistics to determine if a solution is available, or how long the next training cycle might take to complete.