Create or Update Prediction Type (Hub)

[This topic is pre-release documentation and is subject to change.]

Either creates a new Prediction type or updates the type if it already exists.


The request is constructed as follows:

HTTP Verb Request URI
PUT <Hub-endpoint>/manage/model/predictions/<predictionName>?api-version=2017-04-26&locale=<locale-code>

URI Parameters

URI Parameter Required Description
Hub-endpoint Yes The base URL for your Customer Insights Hub
predictionName Yes Name of the prediction to update/create
locale-code No The localization code to apply to any returned properties. If no locale code is specified, then 'en-us' is used as default.

Request Body

A JSON object describing a Prediction type.


The response includes an HTTP status code and a set of response headers. Upon success, the body will contain information about the newly created prediction in the following form:

    "id": <predictionUrl>, 
    "name": <predictionName>, 
    "type": "Microsoft.CustomerInsights/hubs/predictions"

Status Code

  • 200 (OK) - Update request completed successfully.
  • 202 (Accepted) - Create request was accepted.

Type provisioning is typically a long running operation. The initial request will come back with HTTP Accepted (202) with a location header to GET the status of the operation. The properties in the response body will have a field called provisioningState. Eventually performing a GET for the prediction will return a response where provisioningState is set to Succeeded, indicating success.