--- title: Predictive Intelligence API description: The Predictive Intelligence API provides endpoints that predict a field value based on one or more input fields and a trained solution.Predicts an output field value using a specific solution.Returns predictions for multiple solutions. locale: en-US canonical_url: https://www.servicenow.com/docs/r/australia/api-reference/rest-apis/agent-intelligence-api.html release: australia product: REST APIs classification: rest-apis topic_type: concept last_updated: "2026-03-12" reading_time_minutes: 5 breadcrumb: [REST API reference, API reference, API implementation and reference] --- # Predictive Intelligence API The Predictive Intelligence API provides endpoints that predict a field value based on one or more input fields and a trained solution. You can only use this API when the Predictive Intelligence \(com.glide.platform\_ml\) plugin is activated. **Parent Topic:**[REST API reference](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/australia/api-reference/rest-apis/api-rest.md) ## Predictive Intelligence - GET /agent\_intelligence/solution/\{solution\_name\}/prediction Predicts an output field value using a specific solution. ### URL format Default URL: `/api/now/agent_intelligence/solution/{solution_name}/prediction` ### Supported request parameters
| Name | Description |
|---|---|
| solution\_name | Name of solution to use for predictions. For example, `ml_incident_categorization`.Data type: String |
| Name | Description |
|---|---|
| Solution definition input field key-value pair | Name-value pair of the solution input field. For example, enter the name: `short_description` and the value: `Unable to connect to VPN`. Data type: String |
| Header | Description |
|---|---|
| Accept | Data format of the response body. Supported types: **application/json** or **application/xml**. Default: **application/json** |
| Element | Description |
|---|---|
| input | Name-value pairs that were specified as input to the call.Data type: Object |
| output | Predicted outcome values based on the specified solution.Data type: Object ``` "output": { "confidence": Number", "outcome": "String", "threshold": Number" } ``` |
| output.confidence | Estimated precision of the prediction as a percentage. For example, **53.84615375762915**.Data type: Number |
| output.outcome | Prediction output field value. For example, an incident categorization solution would return an incident category such as **inquiry**.Data type: String |
| output.threshold | Value of the configured threshold associated with the prediction. Data type: Number |
| Name | Description |
|---|---|
| api\_version | Optional. Version of the endpoint to access. For example, `v1` or `v2`. Only specify this value to use an endpoint version other than the latest. Data type: String |
| Parameter | Description |
|---|---|
| input\_filter | Required parameter if **input\_table** is used, don't use this parameter with **input\_maps** parameter. Filter to select records on which to run predictions. For example:``` sys_id 0ef47232db801300864adfea5e961912 ``` Data type: String |
| input\_maps | Required unless using **input\_table** parameter. Array of input name-value pairs. For example: ``` [{"short_description":"my email is not working"}, {"short_description":"need help with password"}] ``` Data type: Array of Objects |
| input\_table | Required unless using **input\_maps** parameter. Table name on which you want to run predictions. For example:``` incident ``` Data type: String |
| options | JSON object with optional arguments. For example:``` {"top_n" : 5, "apply_threshold":false} ``` Valid options: - top\_n: Number. If provided, returns the top results, up to the specified number of predictions. - apply\_threshold: Boolean. Checks the threshold value for the solution and applies it to the result set. The threshold value is solution threshold for similarity or class-level threshold for classification. Default value is true. - custom\_results\_filter: String. Similarity solutions only. Specifies the allowed set from which results are returned using an encoded query. Data type: Object |
| solution\_names | Required. Comma-separated list of solution names for which you want to run predictions. For example:``` ml_incident_categorization,ml_incident_assignment ``` Data type: String |
| Header | Description |
|---|---|
| Accept | Data format of the response body. Supported types: **application/json** or **application/xml**. Default: **application/json** |
| Element | Description |
|---|---|
| confidence | Value of the confidence associated with the prediction. For example, 53.84. Data type: Number |
| predictedSysId | The sys\_id of the predicted value. Results can be from any table on which information is being predicted. Data type: String |
| predictedValue | Value representing the prediction result. Data type: String |
| threshold | Value of the configured threshold associated with the prediction. Data type: Number |